EXCAVATIO THREF NO E CHAMBERED TOMBS3 10 .

X.

THE EXCAVATIO THREF NO E NEOLITHIC CHAMBERED TOMBS IN GALLOWAY, 1949. BY STUART PIGGOTT, B.LiTT., F.S.A., F.S.A.SCOT., Professo f Prehistorio r c Archaeolog e Universitth n yi y of Edinburgh, AND T. G. E. POWELL, M.A., F.S.A., Lecturer in Prehistoric Western European Archaeology in the University of Liverpool.

CONTENTS. PAGE INTRODUCTION ...... 103 PAET I.—DESCRIPTIO CAIRNHOLE TH 6 F N10 O Y. EXCAVATION . . S Cairnholy Chambered No. I ..... 106 Cairnholy No. II . . . .123 PART II.—DISCUSSION OF THE CAIRNHOLY EXCAVATIONS . . . 129 PART III.—THE GALLOWAY GROU GLYDE-CARLINGFORF PO 0 D14 TOMB . S PART IV.—EXCAVATION OF THE WHITE CAIRN, BARGRENNAN . . 144 PART V.—DISCUSSION OF THE BARGRENNAN EXCAVATION . . . 151 APPENDICE F ...... A- S 3 15 .

INTRODUCTION.

Thanks to the work of Childe l and Daniel,2 it has been possible within e lasyearn th te tmak o t s e certain basic divisions withi greae nth t group of Western European collective chambered tombs, spread from Iberio t a the Orkneys and beyond, and in genera] the products of Neolithic or early metal-using cultures round abou beginnine th t e seconth f go d millennium B.C. These divisions are based on recurrent peculiarities of tomb-plan and architectural techniques, which can be seen to be consistent within certain geographical areas, and can often be correlated with distinctive forms of and other grave-goods. Two major recognisede groupb n ca s , represente locay db l derivatives d variantan n mani s y area f Iberiao s , Franc d Britainan e thougd an , h typologically degenerate tombs can be confusingly similar in ground-

1 Prehist. Comma, of British Isles (1940), pp. 46 if., summarising earlier papers. 2 Proc. Prehist. Soc., vol. vii. (1941) . 1-49pp , . 104 PROCEEDINGS OF THE SOCIETY, 1948-49.

plans, the primary types seem distinct and recognisable. One such type is that of a more or less rectangular burial vault set at one end of a long cairn—the Gallery-Grave, e otheth a round r an r polygonao d l burial chamber approache narroa y rouna db n wi dt passag se cairn—th d ean e Passage- Grave. e mosOnth f teo notable region f colonisatioo s buildery nb f Galleryso - Graves in Britain is that around the head of the Irish Sea, including Ulster, West and the Isle of Man. In these areas occur a large numbe f chambereo r d tombs (probably betweed 200an ) 0 witn15 h certain features in common, and, from their concentration around Carlingford Loug Northern hi n moute Irelane Clydth th f d Scotlandn ehi o d an , they have been classified as the Clyde-Carlingford group of tombs. The pioneer work of Bryce in the Clyde mouth area, and the later

extensive excavations of Estyn Evans and Da vies in Ulster,1 2 have given faia s r u knowledg materiae th f eo l cultur e builderth f eo f thesso e tombs as represented by grave-goods, and in at least some information on the burial ritual and sequence of funerary deposits. One site has also been excavate Isle t Manalthougf th eo Bu n di . t leasha t eight long in Galloway could be seen from surface indications to represent tombs of the Clyde-Carlingford class, no excavations had taken place in this area. In view of this gap in our knowledge of the Clyde-Carlingford culture, the Research Committee of the Society of Antiquaries of Scotland, when considering post-war plan r excavationfo s n Scotlandi s , gav a ehig h priority to the selected excavation of one or more chambered tombs of this type in Galloway. The two closely adjacent tombs on the farm of Cairnholy, between Gatehous f Flee eo Creetownd an t , Kirkcudbrightshire, were chosen not only for their accessibility but because, although much ruined and plundered, certain features (notably the forecourt area described below) appeared to be intact and likely to provide important information on ritual and the burial sequence. Furthermore, the presence of two sites withie anothe yard0 on 50 nf o sr meant that they coul e excavateb d d conveniently under centralised supervision, and it was hoped that some evidence of the relative date of one to the other might be obtained. Excavation unde e directio th re writer th accordingls f no wa s y carried out between 20th June and 2nd July 1949 by students working under e auspice th e Scottisth f o s h Field Schoo f Archaeologyo l . Owine th o gt remarkably fine weather, which permitte uninterrupten da f eighto n -dru hour workin siteso tw g, e wor daycompleteth s n kswa o d ahea schedulef do ,

1 Proc. Soc. Ant. Scot., vols. xxxvi. (1902), pp. 74-181; xxxvii. (1903), pp. 36-67; xxxviii. (1904), pp. 17-81; xliii. (1909), pp. 337-70. 2 Mainly publishe n Ulsteri d Journ. Arch., N.S., vols. i-v. (1938-42). Consolidated list with bibliographical references in ibid., vol. vi. (1943), p. 4. EXCAVATION OF THREE NEOLITHIC CHAMBERED TOMBS. 105

and a small supplementary excavation was therefore undertaken on a third site described below.1 n additio I e chambereth o t n d e Gallery-Gravcairnth f o s e clasn i s Galloway knows wa t ni , tha isolatee t on ther s dewa exampl a smal f o e l (and typologically late) Passage-Grave withi mile5 1 n f Cairnholyo s e th , CHAMBERED TOMB GALLOWAN I S Y

Pig. 1. (The Cairnholy tombs are Nos. 2 and 3; Bargrennan No. 11.)

eastern of the two White Cairns of Bargrennan.2 It seemed desirable to

Walter M o t r e McCullochthankr du Ou 1e sar ownee th , f Cairnholyo r Andrer Si o Ladd t ,w an y McCulloc Macmillanr Ardwallf M ho o t d tenane an ,th , f Cairnholyo t permissior fo , excavateo nt d an , for every helencouragemend pan t durin e workgth . Financial grants wer ee Societmadth y f eb yo Antiquarie Educatioe th Scotlanf so y b d ndan Committe Kirkcudbrighe th f eo t County Councile Th ' . County Council further co-operate loae huta th f no ,y dwheelbarrowb othed san r tackle whicr fo , e hw were most grateful studentx Si . s workee excavationsth n o d , fro e Universitiemth f Edinburgho s , Glasgow and London, as as Mr Allard Johnson of the West Highland Agricultural College. To all these strenuous and cheerful workers the success of the excavation is ultimately due. Thanks are due to the Forestry Commission, on whose land the White Cairn lies, for permission to excavate2 . This cairn, and those at Cairnholy, are scheduled under the Ancient Monuments Act, 106 PROCEEDING E SOCIETYTH F O S , 1948-49.

investigate this very ruined site in the hope of obtaining some evidence s relationshipoit f lacr o ,relationshipf ko dominane th o t , t Gallery-Graves districte odaysfth o Tw ' .wor k were accordingly unplundere e spenth n o t d chambee parth f o tpassaged an r , with interesting results. The following report is divided into three main sections, dealing with cairno tw t Cairnhola se th thad t Bargrennaan ya t n respectively. Fol- lowing these excavation reports, we have added a short discussion of the Clyde-Carlingford tombs in Galloway as a group, based on field-work e summeth carrien i f 194 t o r dou 9 take conjunction ni n wit e resulthth s excavationse oth f .

PART I.—DESCRIPTION OP THE CAIRNHOLY EXCAVATIONS. CAIRNHOLY CHAMBERED CAIRN No. I. The Site and its Setting (figs. 1 and 2). Thi s companionsit caird an n , Cairnhol hig n , boto yII h e hli groun d

abov valleKirkdale e eth th f yo e Burn1 , which t Wigtowruna a s se int e noth

Fig . Sit2 . e pla Cairnholf no y Chambered Cairn sanI . dII

littla y e Ba over hal southe 400-foo e mila th f Th o et . t contour actually and the excavations were carried out with the permission of the Ancient Monuments Department of the Ministry of Works, through the Inspector of Ancient Monuments for Scotland, Mr Stewart Cruden. 1 Nat. Grid Ref. 517538; O.S. 6-inch sheet Kirkcudbrightshire XLVII S.W., Royal Comm. Anc. Mons. (Scot.), Kirkcudbrightshire Inventory (1914), No. 288. EXCAVATION OF THREE NEOLITHIC CHAMBERED TOMBS. 107

passes throug site Cairnholf hth eo y It liei gentl:n so y sloping land rising fee5 Craigmuln o t62 o t e halnorthe mila fth o et , wit s lonhit g axis along contoure southe th th gentle o th T , . e slope continue fee0 50 t r beforfo s e e steeth pe Kirkdal th clef f o t e Glen e northbegins th looke o on p T , u s. the valley of the Kirkdale Burn between Cairnholy Hill (850 feet) and Cairnharrow (1497 feet), which dominates the northern skyline. South- wards there is a wide view over Wigtown Bay, the peninsula of Wigtown being terminated by the Isle of Whithorn 11 miles away across the entrance of the Bay, and beyond and to the south-east the mountains of the Isle of Man, 35 miles away, are conspicuous on clear days. The solid geology is that of the Llandovery subdivision of the Silurian, with flags, shales, grits and greywackes, overlaid almost entirely by glacial deposits. Along the eastern side of Wigtown Bay the Early Post-Glacial Raised Beach is merged in the spread of estuarine deposits brought down Rivee byth r Cree.1

Visible Structures before Excavation (fig. 3 anid PI. XVIII, 1). The beesitd eha n considerably mutilated, probably mainly withie nth last century or so when the present fields are likely to have been enclosed massive th y b e stone dykes bounthaw no t d themr thesFo e. dykee th s cairn offered an obvious quarry, and indeed one such dyke cut diagonally acros site sth e from south-wes north-easto t t t coulI . seene db , however, that the cairn had originally been approximately 140 feet long by 33 feet wide, that part lying to the north-west of the dyke and its accompanying cart-track being extremely denuded and visible only as a slight swelling pasturee inth . e south-easToth e dyketh f o ,t however e cairnth , , thoug placen hi s robbe ground-levelo dt , still retained many presenfeaturess o t wa p d u t an , a height of 2 feet 6 inches and more in places. Along the southern edge, stones of a revetment could be traced intermittently projecting from the rough turf for a distance of over 100 feet in an approximately straight line, and at the eastern end there were standing the conspicuous stones of a shallow crescentic facade or forecourt, 23 feet wide and with a depth of littla easymmetricalls ovewa feet 5 rt I . y planne consisted dan eighf do t stones, the two at the centre of the arc forming a pair of impressive portal stones standin fee6 g t high abov e cairth e n surface n fronI f thes. o t e wa a stons e leaning obliquely forward. Cairn materiae s pileth wa n ldi forecourt area to a height later ascertained to be 3 feet above the original ground-level. Behind the portal stones lay the chamber, with an overall length,

1 J. Pringle, British Regional Geology, South of Scotland (1948); A. D. Lacaille, Proc. Geol. Assoc., vol. lix. (1948), pp. 151-71. 108 PROCEEDINGS OF THE SOCIETY, 1948-49. -

portals included feet 6 t coulI 1 f .see e o ,d b n tha chambee th t divides rwa d into two parts by a high upright slab, and that the rear compartment, or chamber proper, had been dug out to a great extent. The front com- partment, or antechamber, was, however, full of debris up to the level of denudee th f o dp to cairn e th . There wer roofino en g slab situ,n si though large tabular stones lay in confusion to the south of the chamber on the cairn surface. Method of Excavation. The problems presented by the site resolved themselves into three: to examine what remained of the cairn itself, and to trace its revetment; to examine the forecourt area in relation to probable ritual performances in connection with the burials; and to recover what evidence of burials, sequence of deposit, and grave-goods that the half-ruined chamber might provide. Th efro ) plannes sit m60 base a einc wa : 1 scal a fee 5 1 ho ( o -f dt t e o line running paralle aree lsouth e th caire a wit th d axie th o hnf th t an ,s o comprising the chamber and forecourt were separately planned in detail inc 24): 1 scala fee 2 1 ho ( .o f t t e o Sections were drawrathee th o nt r large shallownes vien th i e f ) winc 1 o footscale : 1 12 1 ( ho th f t eo f o s deposits and the complexity of their make-up. The revetmen examines cuttingso wa t tw n di , exposin gtotaa l length d showinfee0 o4 fan t s relatio git e cair th no nt material outsid s lineeit . A t througsectiocu s hnwa what remaine e side cairth th e o f nt walo d l of the burial chamber, and two further areas within the cairn were stripped down to the old ground surface. The forecourt areexcavates awa halveso tw n di mediaa , l baulk being left unti lata l e excavatioe stagth f eo enablo nt blockine eth g materian i l tha tstudiee areb recorded o at dan sectionn i d whole Th .e area within the arc of the facade, and for a distance 15 feet beyond the portals, "was stripped dow undisturbeo nt d ground-level. The cramped area within the narrow chamber precluded the leaving of standing sections during excavation t thesbu , e wer epapen o buil p ru t excavatioe th s a n proceeded positioe th l find al d f san no , were plotten do to these sections and the plans as they were discovered.

Summary of Results. A chambered tomb for collective burial, architecturally a member of e Clyde-Carlingforth d group distributed over south-west Scotland, Ulster anIsle Manf d th eo bee d ha , n beginnine builth t secone a t th f go d millen- nium B.C. in wooded country on a gentle slope overlooking the sea. Its builders used pottery aki thao nt t foun e Isl f ManUlsten th i eo d d ,an r EXCAVATIO F THRENO E NEOLITHIC CHAMBERED TOMBS9 10 . e Clydth thougn ei moutt hno h area d leaf-shapean , d flint -heads. Pitchstone was, however, imported from Arran furthed an , r trad shows ei n by a large jet bead and a piece of a highly polished ceremonial jadeite . cremationy Buriab e forecours th wa ld e focuan , tth f areo s s wa a elaborate funerary rites, including the making of fires and the deposition of offerings of pottery and edible molluscs, followed by a massive ritual blocking of the entrance to the tomb after its effective use was ended. Before the deposition of this blocking however, at the beginning of e Bronzth e Age e antechambeth , tome uses th furthe r f bwa dfo o r r burials by , accompanie Laty db e Neolithic potter type th ef y o know n fro e habitatiomth n site Glenlucn so e Sands alsd an o, beaker f Typo s , eB and a large flint . Finally, in the Middle the rear chamber was reused for a burial by the side of which a cup-and-ring-marked stone had been set up. DESCRIPTION. The Cairn. lowee Th re cair parth f no t which alone remaine homogeneouf o s dwa s construction wherever examined consisted an , a simpl f o d e pil f largeo e rounded boulders with a minimum admixture of earth. At its lower edge at least there was a rough attempt at containing the mass immediately behind the revetment by obliquely set slabs leaning inwards towards the axial h'ne of the cairn. d land-surfacol e Th e t appeaunde no e cair d th o rhavt rdi n e been artificially prepared t patchebu , f burno s t amorphoud soian l s carbonised wood suggested the lighting of fires, either ritual or more likely as a part preliminare oth f y clearanc woodea f eo d glacia e siteTh . l deposite th f so region would everywhere support woodland, almost certainly oakwood of a modified Q. sessiliflora type similar to that still surviving in patches in Highland Scotland,1 and such land clearance would be necessary. Identi- fiable carbonised wood from the site represented oak (Quercus sp.) and hazel (Corylus Avellana Linn.). (See Mr Orr's report in Appendix D.) Owing to the slope of the site there had been some lateral shifting of the loose cairn material downhill, to the detriment of the revetment, which had been pushed forward and allowed the cairn to slip outwards beyond its line.

The Lateral Revetment (figs. 3 and 4; Pis. XVIII, 2: XIX). For over 100 feet on the south side of the cairn, and for a short stretch at the north-east corner, remains of a revetment of slabs originally set . TansleyG . A . , 1Cf Brit. Islands theird an Vegetation (1939) e soi . 343-9th l pp , d analysean , n i s Appendi. xP 0 11 PROCEEDING E SOCIETYTH F O S , 1948-49.

upright, and running in an approximately straight line, could be traced before excavatione southerth e westerf o th t d nA nen line. , againse th t modern stone dyke cutting across the cairn, a large stone still stood upright, but elsewhere the relatively thin slabs had been tilted outwards by the downhill shift of the cairn, and projected obliquely near the crest of the rough slope formin caire gth n edge singlA . e upright still remained visible nea north-ease th r t cornecairne th f o r, This revetmen examines twa runniny db cuttingo gtw towardn si lins sit e from the bottom of the visible cairn slope, and then extending laterally o expost s outeit e r e cairfaceTh n. material outsid e revetmeneth s wa t studied with particular care to determine whether it represented an accidental tumbl stonesf eo deliberatela r o , y built structur "extra-revetmente th f eo " type known fro Cotswole mth d long cairn elsewhered san . There seemed

no doub1 t to us that it was no more than an accidental spill of cairn material, and indeed the revetment could hardly have collapsed so consistently had it been buttressed by built material outside it. Scattered along the front ground-surface d revetmenoth ol f e th n o t e were numerous lump whitf so e quart varyinf zo g sizes. The slabs formin revetmene gth witp beeu d t htha nsingularlse y shallow stone-holes or none at all, and their subsequent collapse is not surprising. It is probable that they were linked by intermediate dry-walling, but this had suffere suco t d exten n collapse ha th n i t e that only ambiguous traces of it could be seen here and there. On the north side of the cairn, at the top of the slope, the collapse had taken place inwards as the cairn shifted downhill. vien I thwf o e absenc surviviny an f eo g extreme stonesth d an ,e shallow- nes stone-holesf o st considereno s wa t ,di worth while attemptin traco gt e the course of the revetment round the western end of the cairn, beyond the stone dyke. What can be seen suggests the likelihood of a square- ended cairn, with approximately parallel sides d thaan ,t would conform in general terms to the known Clyde-Carlingford plans, especially the Scottish tombs n ArranI . Dunad an , Carn n nBa Bea g 2 appea havo t r e more or less parallel-sided cairns, while Monyquil,3 with remains of its revetment stones visible, seems certainly to have such a plan. The Irish tombs tend to have a more rounded cairn outline, or to approximate to the trapezoid form of Scotland and the Isle of Man. The eastern end of the cairn was examined with care in order to determine the structures between the lateral revetment and the ends of the crescentic facade. But no evidence could be recovered for any form of revetment or walling joining the eastern ends of the two lateral revet-

Arch., vol. Ixxxvi. (1937), p. 136; Proc. Prehist. Soc., vol. v. (1939), p. 125. 12 Proc. Soc. Ant. Scot., vol. xxxvii. (1903); ibid.,36 . p , vol. xliii. (1909) . 343p , . ' Ibid., vol. xxxvii. p. 68. CAIRNHOLY: CHAMBERED CAIRN NQI

UPRIGHT STONE 3™, UNDISTURBED GROUND COLLAPSED JUVETMENT L/.'J EXCAVATED AREAS £3? ROD CA1R.F YO N CAIRN 5PR.EAD

Pig. 3. 112 PROCEEDINGS OF THE SOCIETY, 1948-49.

ments to the north and south tips of the facade setting. In the south, the collapse of the revetment and the slip of cairn material might be thought to have destroyed such structure, but at the north-east corner, where the cairn was on nearly level ground and no slip had taken place, the revet- ment could be traced in progressively diminishing stones as it ran eastwards, t altogetheou e di o t r before cairth f neo materiareachind en e glth which sprea t beyonou d e fagadeth d t bot A e "horns,.th h s wa n facti " t i , exceedingly difficult to make any distinction between the body of the cairn e uppe th d oute d an r ran e forecourlayerth f o s t blocking with whict i h merged imperceptibly.

The Facade and Forecourt (figs. 4, 5, 6; Pis. XX-XXIV). The most conspicuous sitth e f beforo e e excavatioe th s nwa f standing-stoneo c ar s easterit t a scentrao ntw ende l Th member. f o s this setting of eight stones formed the monumental portals of the burial

CAIR.NHOLY N? I ElEVATIO F O FACADN E

Fig. 5.

chamber, flanked to north and south by two groups each of three stones, forming an irregular crescent. Within the area enclosed by this crescent, and extending beyond it, was a filling of cairn material, from which in fact facade th e stones protruded. On excavatio founs wa d t n i tha facade tth e presente dcurioua s mixture of symmetr asymmetryd yan , bot plan hi n ari elevation di . 5) n d (figsan 4 . The southern half was laid out in a shallow but more or less regular curve, and the stones comprising it were graded in height from the southern portal- stone (9 feet above the old ground-level) to the end stone of the arc (6 feet 6 inches). Fragments of dry-stone walling remained between the portal s neighbourit d stoneso d ha southe tw an d th betweed d an n , an -o ,en e nth originally e orthostatslinketh l dal northere Th . nfacade th hal f o fe con- sisted of an almost straight line of stones (one now leaning forward to a considerable degree) which were not graded in height in the manner of Proc. Soc. Ant. Scot.] [VOL. LXXXIII. CAIRNHOLY NQ I- DETAILED PLAN; F CHAMBEO R 8^ FORECOURT

LIMI F EXCAVATIOTO N

APPROXIMATE EDGE OF CAIR.K1 MATERIAL

5TONE-HOLE WITH OVER.

OLD SURFACE A BURNT -

lit

FIND/ ® FOOD-VESSEL 5HERD5 ©'WESTERN' NEOLITHIC 5HER.DS CHER® T LEAF- © 'PETERBOROUGH' NEOLITHIC JHERD5 ® FLINT KNIFE © 'B' BEARER 5HEED5 (7) POLISHED'JADEITE' FRAGMENT PITCHSTON® E FRAGMENT ©'WESTERN'NEOLITHIC BOWL ® PITCHSTONE FLAKE DEPOSITS IN BLOCKING

©SEA-SHELLS © NEOLITHIC BOWL LJt I Al LJ Ur UPRIGHT 5TONE a JET BEAD FORECOURT BLOCKING LEANING -SLAB NEOLITHIC SHERD5

CAIRN 5TOMEJ COLLAP5ED REVETMENT

ROUGH PAVING

S.P. 19+9

[To face p. 112. STUAB . POWET.E . TG PIGGOT . LT D AN T EXCAVATION OF THREE NEOLITHIC CHAMBERED TOMBS. 113

southere th n stones northernmose th , t being very nearl heighe e yth th f o t intermediato tw pai e portalsf o rth d ean , stones lowerwalliny Dr . g sur- vived between the northern portal stone and its neighbour to the north. The stones had evidently been set up in stone-holes from behind the forecourt, as they butted close against the forward edge of their holes with practically no room for packing-blocks, and any ramps, such as would be expected with stone-hole stoner sfo thif so s height, must have been behind and covered by the cairn. All these stones, as those of the cairn revet- ment locae ar , l "whinstone" masses fro glaciae mth l drift. Almost centrally in the forecourt, before excavation, an additional stone coul seee db n leaning outward witd s butsan h it t embedde cairn di n material opposite the entrance between the portals. On excavation this was found to be a large pillar-like stone originally standing upright in a socket made for it in the blocking material, and serving as a closing stone entrance th tombdescribes e i o th t t I o et . greaten di r detail below. As described above, the excavation of the forecourt area was under- taken in two halves, the southern first. In both halves the sequence of deposits was the same, and it was clear that the forecourt had been filled deliberately and as a unit. Beneath the turf small cairn stones mixed with earth were found, extending in a roughly semicircular spread to a maximum distanc somf eo fee3 e1 t fro portae mth l stones mergind an , g almost imper- ceptibly with the rest of the cairn material at the "horns" between the outermost e stonefacadth f o se e beginningsettinth d e lateragan th f o sl revetmen nortn o tsouthd this han A s .materia removes founs wa l wa d t di that it covered a massive arrangement of deliberately built stone slabs, leaning inwards to form a semicircular mass of blocking in front of the entrance betwee portae nth l stones, extendin fee0 1 tr beyongfo d themd an wit hroughla y semicircular outline edgs it ; e returnin meeo g t facad e th t e settin fee0 g1 t fro northe entrance th m fee th sout7 e n o th .td n han eo e earthth n yI cairn material covering this blockin e souts th n hwa go foun a smald l dis f thio c n stone roughly chipped roun e edgesdth , about 3-2 inches in diameter. Another similar disc was found in the blocking norte onth h (fig8)d , Nos.an 9 . 7 Nea. entrance th rsouta e s th n hwa eo worked poin f whit o tspread an e , flind6) . ove t e edg(figNo th rf th, eo 9 . e blocking nea junctios rit n wit facade hth e settinscattesouta e s th n hwa go r of small gritty sherds probabl f Westeryo n Neolithic pottery. They were l abradedal suggesd an , t that thebeed yha n lying expose weathee th o dt r for some time before being covered by the cairn material. The blockin s carefullwa g y removes structurit d an d e studieds It . inner core, immediatel entrance fronn yth i e chamber f o th t o et coms wa ,- posed of small stones and earth, and this was covered by a series of large blocks, many of them flat slabs of a size that one man could just con- veniently lift, piled in overlapping, inclined layers to a total surviving VOL. LXXXIII. 8 4 11 PROCEEDING E SOCIETYTH F O S , 1948-49. height of 3 feet above the undisturbed ground surface of the forecourt. outee Th r edgformes fairlea wa f do y regular serie lonf so g inclined slabs, within which the uppermost part of the inner area of the blocking was made up of tumbled boulders (figs. 4 and 6). The narrow space between the portals (only 1 foot 6 inches wide) was carefully blocked by a series of thin slabs filling the space, with intervening smaller stone placesn si portale fronn th I f . o ts three stonesn o wer t ese ground-leved ol mako e t th s a esill abova d y sucan ,n i lwa e h a thi s e silon l of the slabs blocking the actual entrance-way projected forward beyond the front edges of the portals and into the material of the forecourt blocking. Immediately above this projecting slab there had been contrived a seating for the leaning stone observed before excavations began, and it was clear that this stone, 6 feet in length, had originally stood upright in this socket, s basit a little e ove a footr ground-leved abovol e e blockinth e th n i l g material, and had acted as a closing stone to the narrow space between the portals, which it approximately fitted. The socket was formed by small slabs diagonally wedged in round the base of the closing stone, its main weight being stone taketh ey n b sla b already mentione projectins da g from the remainder of the entrance blocking between the portal stones. In and under the blocking was a series of deposits of pottery and other objects, evidently place position di e blockin th s s beinna gwa g builtr o , before its deposition. Near the northern edge was a mass of sea-shells, closely compacted, and representing two or three handfuls of the shells of winkles, limpets, mussels, and whelks.1 Near the north portal were sherds of a carinated round-based bowl of Neolithic ware (fig. 7, No. 2) and a flake of Arran pitchstone, and near the south portal a single globular bead of jet or lignite (fig. 9, No. 2). Both these finds were in the blocking, and not on the old ground-level. Almost on the central line and at the base of the blocking, 4 feet from the entrance, were a few sherds of Peterborough Neolithic pottery (fig. 8, Nos. 3 and 4). When the blocking material was removed it was found that at five places on the old surface fires had been lit and had burnt fiercely into the soil: where charcoal could be identified it was of oak and hazel. There wa appareno sn t arrangemen thesf to e centrae th n o ,l s thougwa e hon lin feee9 t fro e portalmth d threa lin an sn ei e south-wes f thisd o t an , the fifth away to the north. Near this last hearth it was found that a thin layer of clean earth had been laid down over a sixth hearth, a large

r Stephen'D e e easSe 1 thif th so t o sreport depositr fa Appendin i tt unde d an No , larg a r. xC e slab of the outer edge of the blocking, was found a deposit of vegetable remains comprising the husks of oat (Avena sp.), seeds of Polygonum Persicaria Linn, and probably achenes of Ranunculus sp. Dr Nelson of the Royal Botanic Garden, Edinburgh, who has kindly examined this material, assures us that it is modern, and must represent food-stuff brought down to this position by a mouse or similar salutaranimala s i t I y. warning against accepting apparently ancient deposits without careful checking! Proc. Soc. Ant. Scot.] [VOL. LXXXIII.

CAIRNHOLY NO I - 5ECTION5

CLOSIKiG STONE (FALLEN FORWARD;

.CAI RN • JTONEJ:>.v".?

REVETMENT " (FALLEN INWARDS)

R.OBBED SURFACE OF CA1R.N REVETMENT ^FALLEN OUTWARD5)

D

S.P. 194-9

Kg. 6. STUART PIGGOTT AND T. G. E. POWBLL. [To face p. 114. EXCAVATIO F THRENO E NEOLITHIC CHAMBERED TOMBS5 11 .

oval burnt area north-east of the north portal, completely masking it. Under this same clean earth spread, and partly in the hearth area, were numerous sherd simpla f so e Western Neolithic round-base) 1 . t (figNo d, po 7 . pitchstona d an e flake. The large hearth already e centradescribeth n o e forel s th axida f - o s court was found to lie over, and to penetrate into, the filling of an irregular hole dug in the old ground-level, approximately 4 by 3 feet and having a maximum dept littla f ho e unde filline foot1 r Th g. beneat e burnhth t hearth layer consisted of earth and rather large stones, several of which were partly embedded in the walls of the hole, as if forced there by pressure. The circumstances were entirely suggestive of the socket for a standing- stone, simila o thost r e know n Englani n d from such site s Aveburya s , and nearer to Cairnholy, in the Monument of , West Lothian.1 The features revealed by the excavation of the forecourt of Cairnholy I sugges e followinth t g sequenc f constructioo e d ritual an nn fron I f . o t e crescentith c facade settina cleare s gwa d space upon which five ritual fires were lit: a sixth fire, large and immediately opposite the entrance chambere th o t mads wa , e afte standing-stonea r , whic occupied hha d this positio earln a t yna forecourt'phase th f eo s beed useha n, bodily removed. heartBe yth h neares norte th ho t t porta offerinn a l mads wa g e involving the deposition of broken fragments of a pot (not all sherds of the pot are presente edge th e slightl d ar s an , y abraded a smal d lan ) flake f Arrao n pitchstone. After an interval, suggested by the abraded potsherds, this arecarefulls awa y covere a thi y nb d laye f cleao r n earth, renderine gth surface indistinguishable from remaindee thath f forecourte o t th f ro . Wit assumee hth d completio buriae th f no l rituals withi chambere nth , the area between the two portal stones was built up with slabs and small stones, and the massive blocking in the forecourt area begun, to the accom- paniment of ritual deposits of seafood, potsherds, and a jet bead. During e constructioth e blockinth f no ga seatin s mada closin r gwa fo e g stone that should shut the space between the portals—it is conceivable, though unproven, that this ston originalls ewa y standin stone-hole th e n gi th n eo axial line. With the blocking completely built, and the closing stone standing verticall n placeyi , cairn materia s heapewa l d ove e wholth r e forecourt. The very abraded sherds from the surface of the blocking to e soutth h suggest that thers somwa ee considerable interval before this wa facn si t carried out. This evidence of an elaborate set of ceremonies, culminating in the buildin massiva f go deliberatd ean e blockin e tomth f bgo entrancen i s i , accordance with what is known of the forecourt area of other chambered tombe relativelth n i sinstance w yfe s where this beeha s n examiney b d 1 Proc. Soc. Ant. Scot., vol. Ixxxii. (1947-48) 68-123. pp , . 116 PROCEEDINGS OF THE SOCIETY, 1948-49.

careful excavation. The pioneer work of W. J. Hemp in the Passage-Grave in has been followed up, in the Gallery-Grave group, by Mrs Clifford and W. F. Grimes in the Cotswolds and , and by.Estyn Evans and O. Da vies in Ulster.1 For the Scottish Gallery- Graves we had no evidence until the Cairnholy excavations, which shows that not only does it conform in general terms with the evidence from Ulster and elsewhere, but that in many respects it produces an even greater bulk of materia evidencd an l rituan eo l than many previously excavated sites. This aspect wil discussee b l t lengtda latea n hi r sectio thif no s paper.

The Burial Chambers ; 6 Pis d . (figsXXVan 4 . , XXVI). The constructio buriae th f no l chamber behin portale dth curiou s swa s and unusual. For a length of 6 feet westwards from the rear of the portal stones ran a parallel-sided, roughly rectangular gallery, about 2 feet 6 inches in width and formed by two orthostatic stones on the north, and one, and the stone-hole of a second, on the south. All roofing had been removed, heighe side-stonee th th t f o t bu s shows that .it cannot have been less than 6 feet above the floor. • The western end of this gallery or antechamber was formed by a tall thin slab, risin heigha feeo 4 g t inche f 6 to t s abov floore et betweeth se , n two very massive but lower parallel stones extending westwards, and with another high thin transverse slab forming a rectangular -like chamber 6 fee inche 3 tsam e th f se o lon widtd g feean 2 h ( inches 6 tantechamber e th s )a . There was thus no possibility of communication between the antechamber anchambere lattee dth onlth n d ca ry an ,hav e been entered fro tope mth , before the capstone had been placed in position, or from the side, over one of the low side-stones, after this had been done. The side-stones were singularly large, the southern being 10 feet in length, 1 foot 6 inches thick fee2 inche6 td an northere s th high d an ,n scarcely smaller. No trac f roofineo g remained , e lonexcepon g r narrofo t w slab which had been tipped into the half-filled chamber and which, from the evidence available from the less badly destroyed chamber at Cairnholy II, where the roofin partls gi y interpretee intactb n ca , e corbellin pars th d a f o t g inter- mediate side-stonebetweew lo e capstones th e nth d an s , which must have slab d height a les en t no sa restee f thao tth n feen4 do inche 6 t fee5 to st above the floor of the chamber. While approximatin a plan "segmentedgi o nt " Gallery-Grave plan, the remarkable height of the stone between the two elements of the chamber precludes it from being regarded as anything in the nature of a septal stone. We have to regard the two elements, antechamber and chamber,

Detailed references to these excavations are given on p. 132, where the evidence is more fully discussed1 . EXCAVATION OF THREE NEOLITHIC CHAMBERED TOMBS. 117

as separate entities probleme Th . s raise thiy db s abnormal pla dise nar - cussed at a later stage, and it is sufficient here to note that the only real parallel to the chamber-antechamber plan of Cairnholy I is in Cairnholy II. The antechambe e filles founb wa dro t d with eart d stonea han o t s height of from nearly 3 feet between the portals to 1 foot 6 inches at the western end e uppeTh . r par f thio t s fillin s sterilwa f gf o findo e r o s structure, but at a point later found to be about 9 inches above the floor, a rough pavin r layego f flao r t slabs coul e floo s tracede th b d wa rn O . another such rough-paved area stretching ove wholantechambere e rth th f eo , and betweepavingso tw e darn nth i , k eart d stoneshan , were scattered fragment crematef o s d bone (principalld san y lying upo lowee nth r paving) sherds and other artefacts. At a point 3 feet from the western end of the antechamber, two small bars of stone lay one on top of the other across the width of the paved floor, as if to partition off the western end. Within this low partition, which only rose to a height of about 6 inches and can hardly be dignified with e a septatitl f th eo la scatte slabs f sherdwa ,o r s which included Peter- borough Neolithic and B beaker fragments (fig. 8, Nos. 1 and 2), and a large plano-convex knife of opaque white flint lying flake-stirface uppermost (fig. 9, No. 5). Cremated bone fragments also occurred in this area. In the filling of the stone-hole which held the tall slab that divided the antechamber from the chamber were found a few sherds of Western Neo- lithic potter a cher d tyan leaf-shaped arrow-head. 4) (fig , Nos9 d . an .3 These were found below the pavement-level, and may therefore be regarded representins a g deposits earlier than antechambe e thosth n eo r floor. aree th a n betweeI partitio w portale lo th e d nth snan finds were com- paratively scarce. In and behind the portal blocking were several small sherd f reddiso s h pottery a s piecwela , s f pitchstoria lo e e immediately behin e southerdth n portal stoneflaka f beach-pebbld o e an , e t flintBu . the most remarkable find in this area was a fragment, with its edges much abraded, of a highly polished ceremonial axe of pyroxene jadeite, found on the paved floor just inside the portals (fig. 9, No. 1). Such are well know s strana y find Britainn i s firse t thith tbu s ,si instanc beine on gf eo archaeologicay founan n i d l context s positioIt . , unfortunatelynis , rather ambiguous abrades it r fo , d edges sugges whico t e t thbelongei ax tha e th t d had been broken long before its deposition on the antechamber floor; but at all events it must have arrived there before the final blocking of the tomb was constructed. On the whole, it is likely to belong to the earlier phastomb'e th f eo s use. reae rTh compartment chamber o , r proper beed ha ,n severely plundered: s cist-likit e structure must have attracted immediate attentions oncwa t ei brought to light during the robbing of the cairn, while the less obviously sepulchral antechambe fortunatels wa r y ignored westere Th . n two-thirds 8 11 PROCEEDING E SOCIETYTH F O S , 1948-49.

were entirely disturbed, though occasional scraps of cremated bone occurred in the mixed filling. A large stone was found lying on the floor of the chamber diagonally across its width, roughly partitioning off an irregular area between itself and the eastern end-slab of the chamber. Within this area the lower foot f earto o hs r seemeo d less disturbed than elsewhere addition i d an ,o nt further scrap f cremateo s d bone ssherd w werfe f unornamenteo a es d pottery of Middle Bronze Age Food-vessel fabric. Most unexpected, however, was the discovery that a small stone slab against the south-east corner of the chamber, the top of which had in fact been visible above the surface of the chamber filling before excavation, was not only deliberately propped into its vertical position by small stones at its base, but that it was carved wit largha e " cup-and-ring" symbol, wit hcentraa surp cu l- rounded by six concentric rings, all in firm pocked technique, and two radial lines more lightly incised stone Th e. itsellozenge-shapeds wa f d an , position i t se corners s it n f standino wa e s (pion n .g o XXVI) . t seemsI , therefore, that ther beed eha nsecondara y burial, presumably by inhumation, the bones having been removed by soil acids, and with a Food-vessel pot, inserted into the eastern end of the chamber in Middle Bronze Age times, and that this burial had the unique accompaniment of a cup-and-ring stele set up by its side. While such designs are well known on living rock surfaces and on the capstones of of Food-vessel date, we do not know of another example of such a carving set up on a small individual slab. After this discovery, search was made for further cup-and-ring carvings on stonee monumentth f were o sw ed rewardean , e discoverth y db a f yo very weathered example with four or perhaps five concentric rings on the under-surface of a slab lying near the south side of the antechamber, and missine th perhap f o ge orthostatsson roofine r parth o XXV, L f o t(P g . ,1) example Th e recorded fro e chamberemth d cairn, Mid-Gleniroy b , 1 . No n MacWhite, should be mentioned in this connection.1

LIST OF FINDS.2 1. Pottery. (a) From Forecourt. . (EO1 . 810.) Bowl f o Western Neolithic pottery, represente y aboub d t 100 sherds from under clean earth sprea forecourtn di sherde Th e .sar

on the whole iinabraded, though a few have worn surfaces and edges. 1 Journ. Roy. Soc. Ant. Ireland, vol. Ixxvi. p. 70. " The finds have been deposited in the National Museum of Antiquities, Edinburgh, and the registration number is given in brackets in each instance. EXCAVATIO THREF NO E NEOLITHIC CHAMBERED TOMBS9 11 .

ware Th vers ei y good, with fine micaceou outee sth gritd r surfacan , e is burnished, varying from pale buff to greyish, (fig. 7, No. 1). vessee Th l appear havo st e bee nshalloa w round-based bow inche9 l s in diameter and 4J inches deep, with an everted and very slightly thickene wead an k m roundedri d carination undifferentiaten a s i t I . d type not infrequent in the Western Neolithic cultures, though not known befor n Scotlandi e e typ , Th howeveris e. , representen i d Yorkshire x and in Dorset.2 . (EO2 . 811.) Carinated bowl f Westerno Neolithic pottery, representey b d abou sherd0 7 t s fro e materia mth e forecour th f o l t blockin fronn gi t of the north portal stone. Some of the sherds, especially from the base,

Pig. 7. Western Neolithic pottery, Cairnholy I. 1. Bowl from primary deposit in Forecourt; . Bow2 l from Forecourt blocking. (Both £)

have lost their surface, and the interior has a carbonaceous deposit suggestin r cookinfo e r holdingus go g food. Where present e outeth , r surface is black and smoothed. The grit is fine and micaceous (fig. 7, No. 2). out-turnen a s ha Tht dpo e rim, straight neck, marked carination d hemisphericaan l base; 1\inche6 inched san sm diameteri e th t a r representativa s highi t I . well-knowa f eo n type known from Glenluce Sands and other Scottish sites, but particularly characteristic of the Clyde-Carlingford tomb f Ulster.so 3 3. (EO. 812.) Abraded sherds of gritty ware from upper part of blocking, south-east corner . (DeposiAbou5) . 0 sherds4 No t t , mostly very abrade smalld dan , none larger than 1-6 inche 1-y sb 0 sherd e inchOn ., 0-5 inch thick, retains an unweathered outer surface of pale reddish grey-blaca bufd an f k interior fragmento therd tw e an ,simpla e ar f o s e rounded rim, 0-4 inch thick. Probably Western Neolithic ware. 4. (EO. 813.) Sherds of Peterborough Neolithic ware from below blocking on central axis (Deposi. Abou4) . smal6 No tt l sherd l lesal s s than 1-0 inch across, including fragments probably of rim and shoulder. Grey surface, flaky texture d ornamentean , d with whipped cord "maggot" impressions and blurred lines probably of bird-bone impressions. (fig4) , Nosd .8 an 3 . 1 Proc. PreMst. Soc. (1937), p. 190 (Kemp Howe).

2 Ibid. (1936), p. 229. 3 Proc. Soc. Ant. Scot., vol. Ixiii. (1928-9). 67 . p , 120 PROCEEDING E SOCIETYTH F O S , 1948-49.

(b) From Chamber d Antechamber.an 5. (EO. 814-15.) Sherds of unornamental ware from chamber in front of eup-and-ring stone (Deposi. Abou1) . smal0 2 No tt l sherd l lesal s s than 1-0 inch across. The three largest retain smooth pinkish-buff outer surfaces and black interior, and are of softish fabric without much f Middlo gritd appeae an b , e o t Bronzr Food-vessee Ag e r o l Cinerary Urn type. Other more gritty sherds, including a probably simple rounded rim fragment comparable with No. 3 above, may be Neolithic. 6. (EO. 816.) Rim, fragment of Western Neolithic pottery, greyish-black with fine micaceous grit. From top of stone-hole filling, west end of antechamber (Deposit No. 2). The rim is flattened, and below it the pot appears to have had a globular form (fig. 9, No. 3).

Fig . 8 Cairnhol. . Peterboroug1 : yI h neolithic sherds from chamber . beakeB . 2 ;r sherds from chamber; 3 and 4. Peterborough neolithic sherds from base of Forecourt blocking. (AH £) 7. (BO. 817—18.) Neolithic sherds on paving in antechamber. About 16-18 sherdswhicf o smoots e hha on , h grey-black surfaces with fine micaceous grit and is probably Western Neolithic. The remainder are of coarse flaky 0-o wart 6 p inceu h thick, with large smoothee gritsTh . d outer surfac s grei e o bufft y , with "maggot" impression n twistei s d cord, and the sherds are typical Peterborough Neolithic ware (fig. 8, No. 1). 8. (EO. 819.) B Beaker sherds on paving in antechamber. Eight sherds probably all from one vessel, of pale red surfaced ware with fine micaceous grit. Where the outer surface remains it shows parallel horizontal line hyphenaten si d techniqu. 2) e. (figNo , .8 . (EO9 . 820.) Sherds from blocking between portal stones. miscellaneouA s collection of about 30 small unornamented sherds, one or two gritty fragments being probably Western Neolithic, other f flako s y Peter- borough texture, and two red scraps, possibly Beaker.

2. Stone. 1. (EO. 821.) Leaf arrow-head of black, chert, 1-45 inches long by 0-95 inch maximum width; 0-2 inch maximum thickness, with rounded base meetin straighe gth t obtusn sidea t sa e angl flaked ean botn do h faces. EXCAVATIO F THRENO E NEOLITHIC CHAMBERED TOMBS1 12 .

From stone-hole filling at west end of antechamber, with sherds of Western Neolithic potter. 4) . y No (fig , 9 . This arrow-head is unusual on account of its substance, and the angular form, wit s straighhit t edges, recalls examples fro e Ulstemth r chambered tombs rather than those from Arran, but in neither region there ar e very close parallels. 2. (EO. 822.) Worked point of greyish-white translucent flint, broken at one corner. Maximum dimensions 1-6 inches by 0-9 inch, 0-15 inch thick. This is a simple flake, the under-surface being unworked, with marked ripplinguppee th d r retouchean , d alon l unbrokegal n edgeo st form a blunt-ended, sub-triangular point. From cairn material over- lying blocking on south side of forecourt (fig. 9, No. 6). . (EO3 . 823.) Plano-convex knifef opaquo e matt-surface greyish-white flint, 3-6 inches long by 1-3 inches wide, and 0-3 inch thick. The flat under-surfac e unworke th s thai e f o t d flake except tha e bulth tf bo percussio s beeha n n reduce p show ti y flaking b de s Th incipien. t hinge-fracture, and the convex upper face has rather rough shallow scale-flaking. Trace a ver f o sy narrow ban f lustrdo botn eo h edges suggest it has been slightly used.1 From floor of antechamber on paving, associated with Peterborough Neolithi clasB d s an cBeake r sherd. 5) s . (figNo , .9 This knife falls into the same general class as those from six Clyde- Carlingford tombs in Arran, and those in Ireland from Dunloy, Carnan- bane, Tamnyrankin Wes Creevykeed an t l (the latter bifacially flaked), as wel thas a l t from Cairnhol . II The. yNo y see represenmo t nativa t e Late Neolithic tradition that found full expressio n Food-vessei n l times.2 4. (EO. 824.) Stone disc, broken, but originally c. 3-2 inches diameter and 0-7 inches maximum thickness, with roughly chippe rubbedd dan aedges. From surfac f forecoureo t blockin soutn go . 8) h . (figNo , .9 5. (EO. 825.) Stone disc, 2-95 inche y 2-b s6 inches d 0-2an ,5 inch thick, carefully chipped at the edges and with a "pecked" hollow 0-2 inch diameter on one face only. From forecourt blocking on north (fig. 9, No. 7). Similar stone discs f unknowo , n purpose, have been recorded from chambered tomb Brittanyn i s , Southern England, Wales, Ireland dan Scotland.3 . (EO6 . 826.) fragment f ceremonialo axe, f pyroxeno e jadeite, maximum dimensions 1-7 inche y 1-b s2 inche d 0-7an s inch thick, with abraded edge t remainbu s f higo s h glassy polis botn ho h faces. From floof o r antechamber on paving, between main deposit and the portals (fig. 9, No. 1). The problems raised by this class of axe are discussed in detail in separata e sectio . 137)n(p . . (EO7 . 827.) Flint flake fro mbeach-pebblea , 1-0 inc 0-y hb 5 inch. From floor of antechamber on paving.

1 The very narrow baud of lustre, i|uite distinct from that produced by cutting grass or straw, suggests that the knife has been used on wood (Cf. Ourwen, Antiquity, vol. iv. (1930), pp. 184-6). 2 Cf. Olark in Ant. Journ., vol. xii. (1932), p. 158. 3 Proc. Prehist. Soc., vol. v. (1939), p. 133; Journ. Roy. Soc. Ant. Ireland, vol. Ixviii. (1938), p. 269. 122 PROCEEDING E SOCIETYTH F O S , 1948-49.

Fig. 9. Cairnholyl: 1. Fragment of jadeite axe from chamber; 2. Jet bead from Forecourt blocking; 3. Western neolithic sherd from chamber; 4, Chert arrow-head from chamber; 5. Plano-convex flint knife from chamber; 6. Flint point from Forecourt blocking; 7-8. Stone discs from Forecourt blocking. (All i) EXCAVATION OF THREE NEOLITHIC CHAMBERED TOMBS. 123

8-10. (EO. 828.) Pitchstone flakes and chips from three locations: (a) Behind south portal ston antechambern i e fro) e blocking(6 mth ; , associated with carinated Neolithic pot No. 1; (c) from the clean earth spread in front of the north portal. Thes f Arra o l appea e al nb pitchstono t r indicatd ean e contacts with that region (see Appendix B). 11. (EO. 829.) Bead, of jet or lignite, from blocking of forecourt immediately in front of south portal. This is a rounded form of the Truncated Convex Bicone of Beck's classification, intermediate between his I.C.—la and I.C.—I/.1 The maximum diameter is 0-8 inch and it is 0-65 inch perforatioe longTh . hour-glass ni s shaped centrae th , l diameter being about 0-1 inch and the outer dimensions 0-25 by 0-3 inch, the difference weao beint e r gdu fro . mstrina 2) . gNo , (fig9 . The bead is not of similar type to those from Neolithic contexts in Southern England (Notgrove, , Windmill Hill), whice har flattened ellipsoid forms. Its analogues seem to lie with Bronze Age types, and this point is further discussed below (p. 134). 12. (EO. 830.) Cup-and-ring-marked stone from chamber. A roughly diamond-shaped stone, overall dimensions 20 by 15 inches, and 5-5 inches thick. On the front face is a central cup-mark surrounded by six con- centric circles, the outermost 10-5 inches in diameter, all in "pocked" technique lightle On .y tooled radia a pail lined r an rathe, r more deeply worked, 0-75 inch apart fro n n centre edge a ru m, th th t ea o et angle of approximately 50° to one another. A single cup-mark 1 inch diameter lies outside the circles, 1 inch from the outer edge (PL XXXI, 2).

CAIBNHOLY CHAMBERED. II CAIR . NNo Th s Settingeit Sited an . (fig2) . The chambered cairn next to be described 2 lies 500 feet north of Cairn- generae th d hollan descriptio, yI settine th naturall 6 f no g 10 . givep yn no applies to this site as well. The cairn itself, however, is in a rather peculiar position, as it crowns the top of a small rocky knoll within the 425-foot contour, and on the edge of the present farmyard of Cairnholy Farm. The cairn occupies practically the whole of the flattish top of the knoll, which breaks away in outcrops on the north-west, and the now denuded chamber e survivinon wit s it h g capstone form a conspicuous s landmarks i t I . traditionall tome yth Galdusf bo mythicae th , l Scottish king t thibu ,s honour is also claimed by the Torhousekie in Wigtownshire.3

Visible Structures before Excavation. e cair d beeTh nha n very extensively s nowher robbedwa d ean , more than 2 feet high, and usually less. Its outline was irregular, but roughly 1 Arch., vol. Ixxvii. (1928) 1-76. pp , . 2 Nat. Grid Ref. 517540; 6-inch Cairnholr sheefo s a t . yI Royal Comm. Anc. Mous. (Scot.), Kirk- cudbrightshire Inventory, No. 287. 3 Ibid., Wigtown Inventory, . 531No . 124 PROCEEDING E SOCIETYTH F O S , 1948-49.

oblong, with overall measurement feet0 4 ,f y abou o b swit 0 s lonh7 it t g axis north-eas south-westd an tnorth-ease th t A werd . e exposeen teth d remains of a burial chamber divided into two in precisely the manner of Cairnholy I, 18 feet in overall length with two large, though disparate, portal stones, but no facade or forecourt setting.

Method of Excavation. The very denuded cairn, set partly on solid rock, offered little hope for the identification of a revetment, if such existed, and a single cutting at the south-west end was made to see the nature of its make-up. The areportae fronn th ai f o lt stones decidea n o , d natural .slope strippes wa , d in two unequal halves to a distance of 10 feet beyond the entrance to the chamber attempn a d an , t mad traco et caire eth n edg botn eo h sidese Th . site was planned from an axial base to the same scales as Cairnholy I.

Summary Results.of A small chambered cairn for collective burial, with distinctive features of plan and architecture paralleled only in Cairnholy I, had been built on the top of a small rocky knoll within a short distance of the larger monument. Evidence of the material culture of its builders comprised fragments of ornamented pottery flina , t leaa knif d f arrow-headan e t evidencbu , f eo burial by users of B Beaker pottery was also obtained. There was no evidence of forecourt ritual, but the entrance had been roughly blocked and a closing stone set up against the portals in the manner of that at Cairnholy I. DESCRIPTION. The Cairn (fig. 10). Very little can be said about the vestiges of the cairn. The cutting at the south-west end showed it to have been built of large stones, mainly flattis impossibls hwa slabst i t determino et bu , edgs eit e wit certaintyy han . On each side of the entrance to the chamber similar cairn construction was found, merging almost imperceptibly inte rougoth h blocking except at one or two points where stones appeared to be set to form some break between the two. There was, however, no definite revetment.

The Forecourt Area d 12) an (figs. 1 1 . In front of the portal stones the natural ground surface sloped gradually for a distance of some 5 feet, and then dropped suddenly. There was EXCAVATION OF THREE NEOLITHIC CHAMBERED TOMBS. 125

therefore little space for blocking or for forecourt ritual. Rough slab-built blocking had, however, been constructe morn di lesr same o sth e manner t aCairnhola s , I spreadiny g wit ha roughl y semicircular outlin fee8 e t beyon entrancee depta dth o t onlf ho t fooy1 bu inche,6 t t mosta s d an , vers wa y t difficuli discoveo t t line junctiof th reo n between this blocking and the body of the cairn at each side of the portals. It did appear possible, however traco t , elina changf eo buildn ei , indicating tha caire th t n spread

CAIRNHOLY - CHAMBERED CAIRN N° II

ROCK. OUTC8.OP

REMAIN.! OFCAIK.N

•• UPRIGHT (£g) BOD CAIRF O Y N i?Z UNDtJTUR&EO GROUND [llj EXCAVATED AREAJ ''nirnu

Fig. 10. t eaca t he entranc ou th sid f o e o foret ma ver y shallow V-shaped fore- court. Centrall blockine th larga n yo y gela ston feee6 t long sucn i , hpositioa n to show it to have been a closing stone originally performing the same functio s thaa nn Cairnhol i twholl w no y, I y recumbent n irregulaA . r hollow between the portals, and extending beyond them, may have functioned partly as a seating for the base of this stone when upright. Two small edgstonen o t e swerse e south-ease founth o axiae dt th f lo t line, covered by the blocking but serving no obvious purpose. There were no trace burninf so surface th n go e beneat blockinge hth findso n d . an , 126 PROCEEDINGS OF THE SOCIETY, 1948-49.

The Burial Chambers (figs. 11 and 12). While the plan of the two-part burial chamber at Cairnholy II duplicates the arrangement at Cairnholy I, the architectural problem is solved differ- ently. The two portal stones, now leaning inwards towards one another,

CAIRNHOLY NQ II DETA!L OF CHAMBER, ^ "^V£-ONA______- AND BLOCKING

FN1 D5 © FLINT KNIFE © BEAK.ER. SHER.DS

S.K 19+9

Fig. 11.

are of disparate heights, that to the north-west being 9 feet 6 inches in height othee th rd beinan g onl feey5 t outee highTh r. element antechamberr o , , is formed by two slabs 5 feet apart forming the sides, with the portals t withise e outenth r ends. Thes a greatet eris o no slabt e o rd s height than 3 feet, but in the southern corner of the chamber is a single tall orthosta fee5 t inche8 t s high, implyin t leasa grooa f o tf this elevation above the floor. The inner end of the antechamber is formed, as at Cairn- holy I, by the slab dividing it from the chamber proper, but here only Proc. Soc. Ant. Scot.] [VOL. LXXXIII.

CAIRNHOLY NQ II: 5ECT1 ONJ

ROBBED CHAMBER.

ORIGINAL FLOOR DESTROYED CLOSING 5TONE (FALLEN FORWARD)

WHITE CAIRJV, BARGRENNAN: SECTION 15 • CA-P-STOHEJ^ FEET r~^ METRES PASSAGE

// £ FALLEN SLABS j:g R.OBBED CHAMB —ii'Miu^^^nu.^

RITUAL PIT WITH CREMATION

STUARI DWELL . E . T.G PIGGOT . T D TAN Kg. 12. [To/acep. 126. EXCAVATION OF THREE NEOLITHIC CHAMBERED TOMBS. 127

a little over 3 feet high. The effective floor-space within the antechamber, excluding the portal stones, is about 5 feet square. The chamber beyond still retaine s capstonedit , partly supportey b d massive corbellin north-westere th n go n sidt otherwisebu e restine th n go end slabs. As at Cairnholy I, the side slabs were rather massive and not more than 3 feet high, but the terminal slab rose to a height of 4 feet above the original floor-level. The interior of this chamber had been completely gutted, and digging had continued into the solid subsoil well below the lowee leveth f ro l sidedgee th ef so slabs . Even in the antechamber the soil had been much disturbed. A leaf- shaped flint arrow-hea d sherddan f inciseo s d pottery were foun thin i d s mixed filling, and on a large stone lying on the floor, and possibly in undisturbed soil, was found a plano-convex knife of translucent beach- pebble flint. A small area in the western corner of the antechamber alone seemed undisturbed, partly protecte e overhanth e sidy th b d ef e o gstonth f o e chamber which projected at this point. On a piece of rough paving lay sherd t leasa f so t three B class beakers shora d tan , distance e awath o yt north-eas patca s burnf ho wa t t material, apparentl yheartha .

LIST OP FINDS. 1. Pottery. 1. (EO. 831—32.) Sherds from disturbed filling of antechamber. About 12 small sherds all less than 1-0 inch across, including a simple rim with diagonal dragged line shallon si w incisions belo wa singl e horizontal cord-impressed line. Three other fragments also have shallow incisions. The ware is flaky or vesicular, without much grit, with pale buff surface (fig. 13, No. 1). The decorated sherds at least are likely to be Neolithic, with affinities with the Hebridean or Unstan group of ornamented pots. 2. (EO. 833-38. BeakerB ) sherds from antechamber. Abou smal0 11 t l sherds of Beaker pottery were foun relativela n i d y compact mas rougn o s h pavine south-westh n gi t cornee antechambeth f o r werd an r e almost certainly undisturbed. Three additional sherds ((/) below) were found neare entrancee rth , away fro maie mth n group t leasA vessel.6 t e sar represented follows:—(PIs a , . XXXI. 1) , (a) Numerous sherd f goo o sd war re d e with fine sparse micaceous grit, of a beaker of Class B. Rim, neck, body and base frag- ments are present, but no complete profile can be made. There cordow lo ia sn beloe ornamen th e rim wth d ,an t consistf o s horizontal lines made wit hsharp-toothea d comb probablf yo teeth8 , widely space extendind dan gbaseo t fro m .m ri Inside the rim is a row of slightly oblique lines made with the same . (6) Sherds , ward includinre f e o wit m ghri fine micaceous grita f o , corded beaker of Class B. The ornament is of horizontal lines 128 PROCEEDING E SOCIETYTH F O S , 1948-49.

of fine cord impressions, and there are three rows of similar cord ornament inside the rim. (c) Five sherds f rim o a coars ,e f ,o includin on ef baso d ean o gtw beaker (presumably Class B) with much d grired-browan t n surface. The rim has impressed cord ornament and remains of a heavy cordon. (d) Sherd of body of beaker of Class B, pale buff-red with fine micaceous e ornamengritTh . t consist f firmlo s y impressed hyphenated ornament in horizontal lines and one zone of zigzags. (e) Sherds of red ware with practically no grit and ornamented in the style of (a), but presumably representing a different vessel from (a)-(d). (/) One large and two small sherds of fine reddish ware with no grit and hard smooth exterior. The ornament consists of rows of vertical shallow stab-marks. Probabl beakeB y r somo r e related ware. 2. Stone. (EO. 839.) Leaf arrow-head f opaquo e greyish-white flint, 1-7 inchez lon 0-y gb 9 inch wide 0-d 1an , inch true thickfore th emTh s i pointe . d oval "leaf" form and both faces are delicately flaked. From disturbed soi antechamben i l r fillin. 3) . g No (fig , .13 This form compares well with the arrow-heads from the Clyde- Carlingford tomb Arrann si .

Fig. Caimhol13 . : IyII . Neolithic sherds . Plano-conve2 ; x flint knife . Flin3 ; t arrow-head; 4. Flint . (All from chamber: 4) 2. (EO. 840.) Plano-convex knife of lustrous and translucent greyish-white flint fro ma beach-pebble , 2-5 inches lon y 1-b g2 inches wided an , 0-4 inch maximum thickness. The flat under-surface is that of the unworke e conved th flake d xan , upper surfac s verha ey fine shallow scale-flaking alon edgee remaindee gon th , e surfacth f o r e retainine gth original white cortex of the pebble. Traces of a narrow band of lustre p suggesti nea e mannea th rt n slighi e rus t comparabl thao et t from Cairnholy I. From floor of antechamber, lying on large stone in probably disturbed soil (fig. 13, No. 2). This knif s i comparable e with that from Cairnhol I yalread y described.1 r beach-pebbl1Fo e flint Appendie se , . xB EXCAVATIO F THRENO E NEOLITHIC CHAMBERED TOMBS9 12 .

. (EO3 . 841.) Broken fragment of scraper blue-gref o y mottled flint, 1-2 inches 0-y b 8 inch. scrapine Parth f o t g edge with moderately steep trimming remains. From disturbed soil in antechamber (fig. 13, No. 4). 4. (EO. 842.) Flint chip f blue-greo y flint, 1-0-y 0b 7 inch. From disturbed soi antechamberf o l .

PART II—DISCUSSIO THF NO E CAIRNHOLY EXCAVATIONS.

Since the two cairns at Cairnholy have such evident relationship in the planning and arrangements of their chambers, and to some extent in the method of closing the tombs after the final burials, it is convenient first to discuss certain problems which relate to both monuments in common.

The Structure and Function of the Chambers. o cairntw e s sharTh a epeculia r structural e featurw s a e r notfa o ,s know, precisely paralleled elsewhere in the Clyde-Carlingford area, in whic e innehth e buriar th hal f o lf chambe s constructei r a massiv s a d e closed cis whico t t h access could never have been gained fro e outemth r segmen portald survivine an t Th . g roofin sidee f Cairnholth o g - d an , yII stoneoutee th f buriae rso halth f o fl chambe botn ri h tombs, show that this forepar designes wa t havo dt ehighea r roof tha e reanth r portionf i d an , parallels in other chambered tombs are valid, it was conceived as an ante- chamber and not a passage. The published plans of the unexcavated monuments had suggested that the Cairnholy tombs incorporated a passage element, foreige normath o t nl Clyde-Carlingford type t thibu s; ignored the longitudinal section, which shows at once the closed character of the rear chamber, and the original height of the roof of the antechamber.1 This separation of the rear portion from the antechamber raises several point interestf so . Compariso inevitablns i e wit closee hth d cist chamberr so s in such Irish Gallery-Graves outside the Clyde-Carlingford series as Labbacallee Cork. Co , , Ballyedmonduff Dublin. Co , Carrickardr ,o Sligo,. Co , 2 and possibly also wit closee hth d cists suc thas ha t withi e "rotundanth " behind the chamber in the Notgrove long cairn in the Cotswolds. In fact,

Irise e planstuda th th hf f so yo Carlingfor d tombs strongly suggests that 3 essentially they contain two main structural (and presumably ritual) elements frona : morr to chambee e on subdivision d an r s behind this—this distinction being most marke Dunloe th n di y cairn, wher reae eth r element is a crematorium-trench in the Yorkshire manner.

1 Hawkes s CfMr . ' remark necessite th n so correlatinf yo g plasectiod nan n evidenc Univ.n ei Land. Inst. Arch. Occasional Paper No. 6 (1945), p. 31. * Journ. Roy. Soc. Ant. Ireland, vol. Ixxvii. (1947), pp. 136 ff., with refs. Arch., vol. Ixxxvi. (1937), p. 119.

3 VOL. LXXXIII. 9 130 PROCEEDINGS OF THE SOCIETY, 1948-49.

A practical problem raised by the closing off of the rear portion of the burial chamber in the two Cairnholy cairns is that of the actual mode of introduction of the burials. Once the dry-walling had been built up on each of the side-stones to a height approximating to the end slabs, and this and the capstone covered by the cairn, access would have been impossible sav removiny eb g some cairn materia side-wallingd an l replacind an , g this afte e buria th rbeed ha ln depositet d no withi s i e chambert i nth w No . impossible that something of this nature was in fact done, since there is evidence from several chambered tomb f Passage-Gravo s e type thae th t later burials in the chamber were in fact inserted in more or less this fashion. Lindsay Scot s noteha t d e Continenthi th t Unival,a sn o d t an 1 t leas excavatora o tw t s have noticed similar circumstance recenn si t years— by e LamaloArnath n i l u Passage-Grav n Herault,i e 2 where triangular apertures seem to have been deliberately left between the orthostats of the chambecapstonee th d an rservo t , methods ea f ingrese chambeo s th o st r for the later burials, and by do Paco in the Alapraia rock-cut tombs in Portugal,3 where the central hole in the apex of the vault would seem to have served a like purpose. At Cairnholy, as at Lamalou and Unival, one would have to assume the minimum of cairn material over the chamber— in fact it is likely that the capstones were always intended to be visible, at leas partn i t . This would also explain t Cairnhola , re-use th f , thiyI o e s closed chamber for a Food-vessel burial: it 'would look like a ready-made cist, easily accessible from the side. Whatever the case with the now vanished primary burials Food-vessee th , l deposit wit s carvehit d stone must have been inserted in this manner.

The Forecourt Area Closingand Stone. In both Cairnholy tombs provision was made for closing the entrance *•'" betwee e portanth l stone a massiv y b s e pillabotn i r slabd o hr an site, s it had fallen or been pulled forward. This feature seems not to have been recorded elsewhere in the Clyde-Carlingford group of tombs, but it does e lonoccuth gn i rcair f Tinkinswooo n n Souti d h Wales 4 wheret a s a , ground ol beddeCairnhol s e dth wa surfact n i do , t notCairnd yII a ean s , a - holy I, in the blocking material at a higher level. An important poin stresses i tmonumentae th y db l fa$ade definine gth forecourt area at Cairnholy I, with the accompanying evidence of complex

ritual activitie portalse absence fronn sucy th sth i f an d o h tf an e,facado e 1 Proc. Soc. Ant. Scot., vol. Ixxxii. pp. 1-49. 2 Ampurias, vol. xi. (1949), p.p. 33-45. Dr Arnal discussed this feature with S. P. on the site in 1949. 3 Alapraia I I (1941) , 112 discussiod p. , , an e sit 1947n th ei Pahnelle n no Th . a tombs woule db susceptible of the same interpretation, and the "silo " burials of Acebuchal (Castillo, Vaso Campaniforme, wely relatede lb ma ) PIIV . . « Arch. Camb., 1915, pp. 253-320. EXCAVATION OF THREE NEOLITHIC CHAMBERED TOMBS. 131

, witII n equall. ha No t ya marke evidency d an lac f r ceremoniao k fo e l performance t alla s othen I . r words provisioe th , crescentia f no c facade and an architecturally planned forecourt to the tomb must have been designe settina s da certaio gt n specific funerary ceremonies that demanded suc hbackgrounda r somfo ef I reaso. suco nn h ritua necessarys wa l e th , ceremonial demarcatio e sacreth f no d ares equallwa a y unnecessary—the ceremonies determined the planning of the area at the entrance to the burial chamber. Performed without the appropriate setting the magic would be inoperative, and without the appropriate rites the facade would be meaningles redundantd san . Discussio e significancth f no e forecour th f o e t ritual must therefore b e confine e stonCairnholo t dth s ea closinr , fa excepyI e entranco s gth n i t e play r parta sparallel Fo . s withi e Clyde-Carlingfornth d grou f tombpo s we must turScottiso n o Irelandnt n i e forecour r hth fo ,d sit eha t area been examined befor presene eth t excavations t eleveBu . n Irish examples causee r nb comparativ dfo e data whicn i , e forecourhth t ares beeaha n excavate dstructurae wholl th part n i d r yan o , l features recorde besn dca t be expresse tabla n dfollowi s e a s (se . 132)ep . t wilI seee b l nelevee thath f o tn Irish sites analysed l haval , e some form of cairn sealing in the forecourt (though in no instance so massive or so deliberately constructed as in Cairnholy I), and all have evidence of ritual performances attested by hearths and scattered pottery. Furthermore, no less than six have evidence of a standing-stone, or the socket for one, within the forecourt area—a remarkably high percentage. In its general

feature f forecouro s t ritual, therefore, Cairnhol s quiti yI e typicae th f o l 1 Clyde-Carlingford series, and in less specific terms the evidence from the forecourts of such Cotswold-Severn tombs as , Notgrove or Nympsfield 2 is again in agreement, though here there is no evidence of standing-stone inappositee b t no s y withi, ma e areahowever nt th I . o t , recall the standing-posts originally set at the eastern end of the un- chambered f Thickthoro s n Dow Dorsen ni d Badshoan t a Le t in Surrey t CairnholA .stratigraph e th I y y clearly shows thastonee th t - hole in the forecourt was bereft of its stone and filled level before the blocking wae thresth builen i comparablt t bu ove , it r e Irish sites such evidence t waforthcomingno s e stone th y hav d ma se an ,bee n removey an t a d time. The relationship of deposits on the old ground surface beneath the blocking, to those actually incorporated in it, is discussed further below,

1 Probable additional examples in unexcavated sites are at Ballymakellet, Co. Louth (Louth Arch. Journ., vol . (1941)x . . 60)p Ballymarlagh,d an , Antri. Co , m (Ulster Journ. Arch. (1948) . 15)p , . 2 Arch. Camb., 1915 . 253-320,pp ; 1916 . 239-294,pp ; Arch., vol. Ixxxvi . 119p . ; Proc. Prehist. Sac., vol. iv. (1938), 188-213. The standing-stones at Tinglestone and Lyneham are in the actual body of the cairn, and so not directly comparable, and the same may be said for that at Manio in Brittany (Antiquity, vol . (1937).x . 441-55)pp , . 132 PROCEEDING E SOCIETYTH F O S , 1948-49.

a P . cS o ry TaJ dS a o cS H Site. o a o 1 4 & o % MH P m bo o J3 t-l .3 a J3a cS a t>>p s To 3 s 1 § » O '(S ft 1 s o o « fl m w o o w

Blockin cairr go n sealing in forecourt X X1 X X2 X X3 X X X X X X4 X

Monolith before portal . X X

Monolit centran hi l area of forecourt X X6

Monolith beyond fore- court X

Probable stone-holn i e forecourt X X X6 X

Hearths and/or pottery scatter below blocking X X X X' X X X X X X X X

Prepared clay floor • X X

Pavement . X X X X X X

Notes.—1. No forecourt. Blocking against chamber entrance only. 2. Deposit over pavement not of usual heavy stones. . Cair3 n sealing take sclad forre y f magainso t portals. 4. Excavation report most unsatisfactory. t certainNo smal. A 5 .forecourt e lsid th e stonf eon o o et . 6. The excavator suggests a ritual pit, but a stone-hole seems more likely. 7. Heartpotteryo n t hbu . References.—Ballyalton, PBNHPS (1933-4), p. 79; Ballyedmond, UJA (1938), p. 49; Ballynamona JRSAI (1938), p 260; Browndod, PBNHPS (1934-5), p. 70; Olady Halliday, PBNHPS (1935-6), p. 76; Olontygora, Large, PBNHPS (1936-7), p. 20; Dunloy A (1938); GowardUJ 59 , . p , , PBNHPS (1931-2) ; Hangin41 . p , g Thorn, PBNHPS (1936-7) ; Legland43 . p , , PBNHPS (1939-40) . 16.p , ; Mourne Park, PBNHPS (1937-8), p. 18. Abbreviations.—PBNHPS—Proc. Belfast Nat. Hist, and Phil. Soc.; UJA — Ulster Journ. Archceology; JRSAI—Journ. Roy. Soc. Ant. Ireland. EXCAVATIO THREF NO E NEOLITHIC CHAMBERED TOMBS3 13 .

but a feature common, to all forecourt deposits examined is the presence of pottery in fragments, in circumstances implying the placing of broken sherd s rituaa s l deposits rather tha e shatterinnth a complet f go e vessel situ.n i This beeha s n commented upon more than onc Irelandn ei d an , again in the Cotswolds, and was explicit at Cairnholy I. As we shaU see, this deposition of groups of sherds, rather than whole pots, is characteristic of ritual within the burial chambers as weD. as in the forecourts in many tombs.

The Sequence Burials,of theirand Relation Blocking.the to The interpretatio e forecourth f o n t blockin t s Cairnholga it d an , yI related phenomena such as hearths and ritual deposits, cannot be under- taken without referenc e sequencth o t e f burialaccessible eo th n si e ante- chamber elemen e tombe presencth f Th o a .tcoupl f o ef sherdo e f o s Western Neolithic pottery, together wit hleaf-shapea d arrow-heade th n i , space betwee stone-hole e edg th e stonnth f th eo ed itselean f formine gth e antechamberbacth f o k , strongly suggests that thes facn i e t represent the remnant of early grave-goods, accidentally preserved in this crack whee antechambe th e res f nth o t r floocleares wa rperhapd dan s repaved. On this paving, as we have seen, were deposited burials accompanied by a plano-convex flint knife, and sherds of Peterborough Neolithic ware and of B beaker; nearer the portals was the fragment of a ceremonial jadeite importans i axe.t I noto t e that bot Peterborouge hth h sherd thosd san e

of th1 e B beaker were deposited as fragments, and in neither instance can have represented complete pots broken in the burial chamber: the rite is distinctively that of the collective tomb builders and contrasts sharply with Beaker Period reus whicn ei individuan ha l burial woul insertee db d with its accompanying complete pot (as for instance at Nether Largie in ArgyU). At Cairnholy II, also, the fragments representing five beakers were deposited as sherds, and in both tombs we have the final burials made accordancn i e wit traditionae hth l ritua brokef lo n sherds, even though some of these were from pots whose origin wite sli h peopl aliene th f eo , single-grave rite. At Cairnhol e deposit th forecoure dividee yI th b n i sn ca dt into otw groups, the first represented by the hearth and sherds of a shallow Neolithic bowl fign (Deposii ) nea .9 4 norte . th r No ht portal stone. This deposit was carefully covered over by a layer of compact clean earth, on which the lower part of the blocking was constructed. This deposit, then, can be considere s primaryda , and inferency ,b t least a ewoule on , d connect i t

possibilite Th 1 bees yha n considere whetheo t s da r this fragment migh t havno t e been placeo dt fulfil the role played by the complete axes in the entrances to chambers at Dunloy (Ulster Journ. Arch. (1938), p. 59) and Creevykeel (Journ. Boy. Soc. Ant. Ireland, vol. Ixix. (1939), p. 53), but the evidence inconclusives i . 134 PROCEEDING E SOCIETYTH F O S , 1948-49.

with the assumed early burials, fragments of whose grave-goods survived antechambere bace crace th ith nf th k o t k a . e Alremaininth l g deposits (sherd f Westero s n Neolithic pottery, including e carinatethosth f eo d bowl nea t beadportale je e th r e th ; th ; edible molluscs; the stone discs) must be of the same general date as the blocking itself thid s fixean i s, relation i de buria th o nt l deposite th y b s presenc f Peterborougo e h Neolithi e axiath cn l o sherd s basd it an t ea s line (Deposit 4 in the forecourt, fig. 4). The blocking, then, was not built until after the burials with Peterborough sherds were placed in the ante- chamber: since these were also accompanie By db beake r sherds blocke th , - s containeit d ingan d deposit earlieo n e rar s tha arrivae nth Gallowan i l y of makers of B class beakers. There is no inherent difficulty in accepting this relatively late date for any of the finds. The jet bead is not com- parable in type with those from Neolithic contexts (Windmill Hill, Hembury, Nympsfield, Notgrove a rough-ou d an , t from Maides nit Castle)n i t bu , rounded biconical for s mrelatei e Earlo typedt th r f eveyo o s n Middle Bronze Age. There is no doubt, on the Irish evidence, that the carinated bowls simila thao rt t fro blockine mth g nea norte rth h porta t Cairnholla yI continued in use until well into the local Early Bronze Age. At both the Cairnholy tombs, then, there is evidence of at least two periods of use of the outer section of the burial chambers, the first phase I bein . ingNo indicate e Westerth y db n rim-sher lead dan f arrow-head, incisee th y db I NeolithiI . No n ci sherdd an s (and possibl leae yth f arrow- head and plano-convex knife from, the disturbed filling). The second phase at No. II is represented by the sherds of five different B beakers, none complete, on a paving which is probably secondary, and at No. I by the intact deposit of a plano-convex knife, Peterborough Neolithic and B beaker sherds, and probably the fragment of a jadeite axe found some distance away. At cairn No. I the two phases are also represented in the forecourt deposits, absent in No. II—the first phase comprising one hearth t leasta , with sherd Westera f o s n Neolithic bowl l covereal , layea y b dr of clean earth, and the second phase by the massive blocking with its con- tained deposit probabld san y hearthe somth f estone-holo e sa Th . e (with the stone1 it temporarily held) would also belong to the first phase. An important conclusion emerges from the consideration of the burial and blocking sequence at Cairnholy I—ritual deposits had been made in e forecourth morn o t e thaoccasione non , and reasonabla y ,b e inference, at the time of each successive burial in the antechamber, and the blocking wae finalth s , irrevocable sealine tombth f o g, after whic o furthen h r burials would be made in it. We cannot regard the Cairnholy blocking removes a replaced dan t everda y funeral, though clearl closine yth g stone here (an dt suca h other site s Tinkinswooda s ) might very well serva s ea temporary mean shuttinf so g off acceschambee th o st r between ceremonies. EXCAVATIO THREF NO E NEOLITHIC CHAMBERED TOMBS5 13 .

But during the effective use of the tomb for collective burial, the forecourt must have bee n opea n n are whicn i a h ritual d firean s t li coul e b d appropriate offerings made. The evidence from the Irish tombs is in no instance sufficiently precise to allow such a sequence to be demonstrated with stratigraphical certainty, though, on the other hand, it nowhere conflicts with such an explana- tion. But recent work in Scandinavia has shown that exactly the same phenomena as at Cairnholy I can be observed in Passage-Graves in Swede Denmarkd onln an interpretee ca yb d an same , th en di manner. e presencTh f quantitieo e f sherdo s s outsid e entranceth e f certaio s n Scandinavian collective tombs (sometimes amountin 40,00o gt r 50,000o 0 fragments!) has been known for many years, and usually interpreted as representing the debris of earlier grave-goods swept out of the chamber in successive clearings before the final burials, since a. long chronological rang potterf eo oftes yha n been present recene Th .t excavation toma f so b near Horsens Jutlandn i , Gr0nh0je th , , have, however reconsideraa o t d le , - tion of the whole problem, for here were not only a mass of forecourt deposits (including complete pots), but also the intact contents of the burial chamber, with abundant grave-goods e chronologicaTh . l rang f boto e h groups was identical, arid spanned the entire Danish Passage-Grave sequence onls potteryywa n i t possibli d an ,regar o et forecoure dth t pottery, buried beneath a massive blocking, as being a series of offerings made during the e Gr0nh0periolighe th tombe th f th f effectivo t dn o f I o j. evidencee us e , Thorvildsenx has reinterpreted the results of excavations at such tombs as Jersdal and Gillhdg in Denmark, and Vestra Hoby and Storegard in Sweden, showd an 2interpretatio w n ne tha e th te forecour th f no t deposits is more satisfactory than the old. Cairnholy I has therefore contributed important evidence to the problem of European collective tomb ritual at large. Relation otherto Cairnholyof II Scottishand I Tombs. The generic relationshi Cairnholo tw e th yf p o tomb otheo st r members of the Clyde-Carlingford group in such areas as Arran and Bute, and again in Ireland and the Isle of Man, has been noted and discussed since their first recognition as a class. It is not the place here to deal with the

question f typologicao s l developmen e tomb-plansth 3 f o e presenct th d an , e or absence of facades and forecourts, though it may be said in passing that the evidence seems to us to suggest that tombs with a shallow crescentic forecour probabld an t y wit t morhno e tha segmento ntw e th f so 1 Aarbeger 1946. 74 . ,p 2 Jersdal—Fra Nat. Mus. Arbedjdsm. 1933, p. 9; Gillh6g—Ark. Stud, tillagn. H.K.H. Kronprins Adolf 1932, p. 36; Vestra Hoby—Med. Lunds V.H.M. 1936, p. 28; Storegard—Med. Lunds Univ. Hist. Mus. 1930, p. 54. 3 As, for instance, by Childe in Scot. Geog. Mag., vol. i. (1834), p. 18. 136 PROCEEDING E SOCIETYTH F O S , 1948-49.

chamber are likely to be early in any such sequence. The Galloway group of Clyde-Carlingford tombs is discussed as a unit below, and in certain feature f architecturo s e they stand apart from e Clydthosth n eei mouth e grave-goodth area t bu , s fro e Cairnholmth y tombse otheth n r o ,hand , provide explicit evidenc relationshipf eo t leasa , tradey b t , between settlers round the mouth of the Cree and communities building the chambered tombs in Arran and Bute. The plano-convex from the two Cairn- holy sites have close parallels from no less than five Arran tombs (Giant's Grave, Dunan Mor, Tormore, Sliddery and Torlin), and even more to the point is the presence of actual Arran pitchstone in Cairnholy I, itself found fivn i e Arran tombs (East Bennan Banm Ca , , Dunan Beag, Dunar nMo Butn i o d Monamoree tw an (Glecknaba n i d an ) d Michael'ean s Grave 1). So far as these two items go, then, contact between the Galloway and the Clyde-mouth groups is satisfactorily established. The pitchstone, petro- logically confirmed as most probably of Arran origin, is an interesting addition to the west Scottish finds discussed by Mr Ludovic Mann.

The carinated bowl from the blocking near the north portaa l stone tells another story. Vessel thif so s typ noticeable ear y absenArrae th n ni t Butd an e tombs, though abundant Clyde-Carlingforn i d tomb Ulstern i s , a derivativ n i d an e site (Mul le Isl Hillf Man.th e typo n i )Th e 3 seems ultimatel Yorkshirf o e b o yt e Neolithic derivation founs i d habitan di an , - tion sites in North-western Britain at Ehenside Tarn, Cumberland,4 and on Glenluce Sand snot far from Cairnholy, and also in Ulster at Lyles

Hill and elsewhere5 . As we have seen, this type of pottery had a long life in the northern Irish Neolithic, and the recent recognition of the associated stone axe t Ehensidsa e Tar products na Langdale th f so e Pike axe-factory confirms the impression of a relatively late date for this ware in the north- west, since this factory flourished in Late Neolithic times. Its presence at

Cairnholy shows tha Gallowae tth y colonie chamberef so d tomb builderd s6 ha their contacts with regione thosth f eo s immediately sout south-westd han , as well as with those in the islands in the mouth of the Clyde to the north. The Peterborough sherds from Cairnholy are the first to be found in a Scottish chambered tomb t theibu , r presenc quantitn i e t Glenlucya e shows that a coastather s ewa l settlemen f thio t s grou f Neolithipo c folk r frofa m t Wigtowno n Bay e inciseTh . d scraps from Cairnhole ar I I y less easy to parallel, but they do seem to have affinities with Hebridean type wareth ef so know n from chambered tombpottersa d san ' workshop in , and if so would suggest extended traffic up and down the

western sea-ways—a7 n inherently probable situation.

1 Bryc Proc.n ei Soc. Ant. Scot., vols. xxxvi—xxxviii xliii.d an . , passim.

3 2 Proc. Soc. Ant. Scot., vol. lii. (1917-18) . 140-149pp , . Antiq. Journ., vol. xii. . 146,p .

4 Arch. Journ., vol. Ixxxviii. (1932) . 96 . p , « Proc. Soc. Ant. Scot., (vol. Ixlii. . 1929)67 . p , 8 Proc. Prehist. Soc., vol. xv. (1949), p. 16. , Proc.. Cf ' Soc. Ant. Scot., vol. Ixix . 480p . ; vol. Ixxxii . 1-49pp . . EXCAVATION OF THREE NEOLITHIC CHAMBERED TOMBS. 137

The fragment of a ceremonial axe of pyroxene jadeite found on the paving of the antechamber of Cairnholy I is of great interest. Although about forty such axes are known from the British Isles, this is the first to be recorded in any archseological context, and although only a small fragment, it is recognisable both by its petrology and its section, with characteristic high glassy e facespolisth s representina n , o h a gquit e typical example. Unfortunately it is not possible to associate it directly with either the earlier Western Neolithic burials in the tomb, or with the later Peterborough-Beaker use, thoug lattee hth r does provid eterminusa ad s depositionquernit r fo s worvien it I f .w no condition earliee th , r date, with implied contemporaneity wit e e buildinb th he tomb o th t f s i o g, preferred. petrologe Th rock"jadeitee f yo th f so " clas discusses si Campbelr D y db l Smith in Appendix A, and he shows that natural sources for some axes under discussion do not occur in the British Isles. On this ground alone, therefore, we are led to regard them as, in the main, imports from con- tinental Europe axe-blades e fore th thith d f m o y borns ,an i b t eou , whichs i comparable with those from Breton and other finds of votive deposits. The ceremonial nature of the axes is hardly to be doubted in view of their extreme thinness and delicacy of finish. e distributioTh f sucno h axe Britain i s s showni , base 14 fign ni n . do data kindly put at our disposal by Miss L. F. Chitty, and summarised in a list included in Appendix A. An earlier map of hers was published by Si r1933,n i Cyri t thix 1bu Fo sl included several axe t strictl type sno th ef yo under discussion, so in the new map these have been omitted, but a few additional examples from other localities have been added. The distri- bution pattern presents several point interestf so Cairnhole Th . y fragment falls into place as an item in a local concentration comprising two axes from Grlenjorrie, Glenluce, and one each from Castle Douglas, Dumfries, and Glencrutchery, Isle of Man, and evidently the result of movements of trade and migration on the western seaways. The axe from the River Spean, Fort William, continue line Wester e tradf sth e th o p eu n Approaches; those from and Sutherland show its ultimate prolongation, and that from Daviot, , use of the Great Glen route. The remaining Scottish axes, from Aberfeldy, Rattray, Stirling and Dunfermline, and presumably also those from Berwickshire and Cunzierton, are probably mor resulte eth likel e tradf b so o yt e inland fro wese mth t rather than from sea contacts on the east coast. This explanation, too, seems probabl e grouth r pefo from Lyme Park, , and Brierlow and Hopton in , and one may plausibly connect these axes with such chambered tombs as in Cheshire d Fivan ,e Well d othean s r site n Derbyshirei s l indicatinal , g 1 Proc. Prehist. Soc. E. Anglia, vol. vii (1933), PI. V, fig. 6 6, opp. p. 156. 138 PROCEEDING S E SOCIETYOTH F , 1948-49.

settlement from the west coast. Indeed, it seems reasonable to associate all these axes with the same seaborne traffic at the beginning of the second

CEREMONIAL AXE LOCATED • 'JADEITE' AXEJ COUNTY ONLY KNOW9 N

BASED OHL.fCHlTTY

Fig. (Fo14 . r gazetteer Appendie se , . 155.p , xA )

millennium B.C. as that responsible for the Clyde-Carlingford tombs and their analogues. The south-west English group, scattered from Cornwall o Southamptot n Water, presumably relate o up-Channet s l trade, again EXCAVATIO THREF NO E NEOLITHIC CHAMBERED TOMBS9 13 .

fro maie mth n -western routes,whole th d e 1seriean s (includin sporadie gth c Irish finds) are likely to be connected with or derived from the well-known Breton ceremonial axes, made from local e "jadeiterockth f so " groud pan foun mann di y chambered tomb votivs sa e deposits.2 e marketh t dBu concentratio f axeno Easn i s t Anglia (five from sites close to Cambridge), and those from the Thames and Canterbury, seem less likel o derivyt e fro ma Breto n source n additioI e natura. th o nt l deposit f jadeito s d chloromelanitean Brittanyn ei e abundancth d an , f eo ceremonial axe f theso s e e samrockth en i s area t musi , rememberee b t d that the same rocks also occur in Western Switzerland and the adjacent parts of Italy, while the allied rock nephrite is found in many regions stretching from north-west Italy northward centrao t s easterd an l n Ger- many.3 Furthermore, finds of axes made of these stones are known from Vosgese th , Thuringi nead aan r Mainz,spreae th sucd f do 4 an h Breton types of axe westwards is to be connected with the occurrence of large rings of jadeite and allied stones, again a Breton type, in Alsace and the Upper Rhine, where they are associated with Grand Pressigny flint blades.5 Our East Anglian axes may then have a Rhineland origin.

The Re-use of the Tomb in Food-Vessel Times. The problems raised by the difficulty of access to the rear chamber of Cairnhol yI hav e been touche cleas i aboven t i dro t tha secondare bu , th t y burial with the stele and Food-Vessel fragments must have been inserted removiny b g dry-stone walling e reafro th side rmf th eo chamber . There is reason to associate the carvings of the cup-and-ring class with the Food- Vessel culture of the Scottish Middle Bronze Age, and in Galloway such carvings, normally on living rock surfaces, form three main concentrations —one near Kirkcudbright ,example w wite wesfe e hDe a th te n so sidth f eo estuary; another westwards of the estuary of the Fleet, including Cairn- thira holy d dan ;mor e sparsely scattered grouWigtowe th n pi n peninsula

south of the Bladenoch (fig. 15).6 A single outlying example has been noted 1 Site se chambere sucth s ha d long e Grecair th r Colt yf no He Dorsetn Mar i sd an e , wits hit shallow crescentic significan e fa9adeb y ma , thin i t s context (Proc. Dorset N.H. Arch.d an Soc., vol. Ixvii. p. 30). 2 C. D. Porde, " On the Use of Greenstone ... in the Megalithic Culture of Brittany," Journ. Roy. Anthrop. Inst., vol. Ix. (1930), pp. 211-34. 3 J. Andree, "Das Natiirliche Vorkommen v. Nephrit u. Jadeit in Europa," Mannus-Bibliothek, (1922)2 2 . . 25-9pp No , . * Prdhist. Zeitschrift, vol. iv. (1912), p. 231; Lindenschmidt, Altert. uns. heid. Vorseit, vol. i. i, pi. I. archaeologicae Inth l museu jadeit o mt Nanca tw e ey ar axe Bretof so n types vere on y, thifind enan from Raon-sur-Plaine othere th d , an smaller, , Donone fro L fooe mf th o te ,northere th botth f o t ha d nen Vosges Mountains. There is a fragment of the butt of what appears to be another similar axe from Marlenheim, Alsace. (Archseological Museum, Strasbourg . 293 . Augus) No P ,6 Notea. . S y tdb 1950. W. Kimmig, "Probleme der jungeren Steinzeit an Hoch-und Oberrhein", Jahrb. d. Schweiz.

Gesellsch.6 f. UrgeschicUe, vol. xl. (1950), pp. 137-155. Known examples are listed in the Royal Commission's Inventories for Galloway, but there are 6 certainly more unrecorded adde e siteb systematiy o st db c field-work. 140 PROCEEDINGS OF THE SOCIETY, 1948-49.

Mid-Gleniroe th f o o e nstona on f eo n chambered cairnWatee th n Lucef so ro . diamond-shapea f o The eus d stele buria e deliberatelth n i l chambep u t yse r t Cairnhola y seemI uniquee b o st . It is likely that the concentrations of cup-and-ring carvings in the regions mentione drelate e locae abovb th y lo d t eexploitatioma coppef no r d othean r ore Bronzn i s timese Ag e : certainly suc ha correlatio n seems probabl Irelandn ei othe n i d r an ,area Nortf so h Britain.1

• CUP-&.-R1NG CARVING *• FOOD-VES5EL

. PigCup-and-ring15 . carving Food-vesseld san Gallowayn si .

PART III—THE GALLOWAY GROUP OF CLYDE-CARLINGFORD TOMBS. The existenc Kirkcudbrightshirn ei Wigtownshird ean t leasa f eo t nine chambered cairns of the Clyde-Carlingford class makes a brief review of groue wholth a s pa e desirabl rouno et d ofCairnhole fth y evidence. Childe has already indicated the main lines of colonisation in South-western Scotland in Neolithic times,2 represented by the distribution of such tombs, but field-work during the period of the Cairnholy excavations has supple- mented the evidence available in his paper and in the Inventories of the Royal Commissio Ancienn no t Monuments. The tombs fall into two distinct groups, one on the northern shore of Wigtown Bay and stretching inland up the valley of the Cree, and the other on the Water of Luce with an outlier near the coast on the east side of Luce Bay, and a probable second in the high ground between the Water of Luc Locd an e h Ryan detaild (figneare an li 1 .list.)n r jus l i so , Al t .

1 MacWhite in Journ. Roy. Soc. Ant. Ireland, vol. Ixxvi. (1946), p. 59 Mid-Gleniroe . Th n example is recorde dfirse herth tr timeefo . 2 Scot. Geog. Mag., vol. i. (1934), pp. 18-25. EXCAVATIO F THRENO E NEOLITHIC CHAMBERED TOMBS1 14 .

within, the 500-foot contour, except for the burial-chamber ruins (no. 1), e Higth hd Gillespian e cair lowen o (no, r7) . land nea e coaste th r Th . sites are usually slopes facing south or south-west. As Childe pointed out, the distribution of the tombs indicates colonisation from the sea, and the utilisation of alluvial stretches along the rivers for cultivation, but most if not all the sites must have been originally in woodland.1 An interesting point is the occurrence of tombs in closely adjacent pairs e map).th f t Cairnholo a 9 d d Mid-Gleniroan yan 8 d an , 3 n d (nosan 2 . The feature is not unknown elsewhere in the Clyde-Carlingford area—one may instance Dunan Beag and Dunan Mor in Arran, and half a dozen example Irelandn i s : curiously enoug t e hrecuri Cotswold-Sever th n i s n grou f chamberepo d cairns (Poles Wood Sout d Northhan ; Eyford an d Newclose; the two Ffostyll cairns for instance) and in the unchambered long barrow Wessef so x (Thickthor Gussagd nan e Dow Dorsetn ni ; Barton Stacey in Hampshire, and others).2 Typologically, the tombs fall into three classes, with a fourth group of indeterminate burial chamber t Newtosa Aule Th d d nWife' an (no ) 1 .s Grave (no. 12). The first class is that of long cairns with terminal chambers, all with crescentic forecourt settings except Cairnholy II (no. 3), the remainder comprising Cairnhol yI (noBorelan, 2) . d (noMid-Gleniro, 4) . nI (no. 9) and Mid-Gleniron II (no. 8)—the forecourt setting of the last is here recorded for the first time. Details of these and the other cairns are . 143pland p e lisan given ,uniforo th o t st n ni m scale givee ar s figsn ni . . 17 d 16an It will be seen that Mid-Gleniron I has, in addition to chambers along the axial line behind the forecourt, one at least at right angles, with a presumed entrance e cairnfro th e thid sid mf th an so e, feature linke sth site wit e seconth h d class f lateral-chambereo , d cairns t a Dranando, w (no. 5), High Gillespie (no. 7) and the of Kilhern (no. 10). Of these, Dranando bees wha n excavated t producebu , artifactso dn interestinn a : g point established f incheverslabsw o w fe e ylo ,a sus , howevere th s wa , high s septalsa , , recallin possible gth y similar practis antechambee th n ei r at Cairnholy I. In this site the chambers had lost their functional entrances unexcavatee casthrouge th th n ei e b cairnedgee e hy thith th d f dso an ,ma sites as well. The relative remoteness of Dranandow and the Caves of Kilhern from the coast carries the implication that these typologically degenerate tombs may be the products of communities later in establish- ment tha primare nth y coastal settlements t Higbu , h Gillespie lies closo et the shore.

1 Cf. Grimes' remarks in Arch. Camb. (1936), pp. 259-82, on the chambered tombs of the Black Mountains: manobservations hi f yo relevane sar Gallowae th o t y tombs under discussion. Crawford, Long Barrows Cotswoldse oth f (1925) . 12p , ; Grimes, loc. p ocit.;f NeolithicMa Wessex

(1932)2 . 142 PROCEEDINGS OF THE SOCIETY, 1948-49.

"•-•~.^>£ yjf

LONG CA1R.N GALLOWAN I J Y

Pig. Plan16 . chamberef so d long cairn Gallowayn si .

O I 2 3 4- S 10 15 - - ~- Fig . Plan17 . f chambereso d long cairn Gallowayn si . EXCAVATION OF THREE NEOLITHIC CHAMBERED TOMBS. 143

sequence Th tomb-plansf eo , wit break-dowe hth chambere th f no s into a numbe f laterao r l units s parallelei , n suci d h site s Cufa s i Hilld an , probably the cairn at Darvel, , and an allied form of multiple- chambered tomb occurs and is described below. But the Galloway series does recall that of the Severn-Cotswold chambered cairns to a remarkable degree, wher evera y similar multiplicatio lateraf no l chambers a reductio d an importancn i n s see i widee typologicalln d i nth en f ro e y later examples.1 thire Th d clas f Clyde-Carlingforo s d tomb Gallowan i s s representeyi d singla y b e example, tha f Cairnderro t y (no . Thi6) . s large roundish cairn has three radially set chambers, which relate it to Glecknabae in Bute and Dunan Mor in Arran, the former (with two chambers surviving) pro- ducing Neolithi Beaked can r sherds pitchstond an , ee latte flakesth d r an , pitchstone and a plano-convex knife. There seems no doubt, therefore, that these three sites belong to the Clyde-Carlingford series: Cairnderry is remote fro t mighe mouti e Creem th t th f bu ,tho well represent inland penetratio frot nno m this coast frot Ballantrae bu ,m th e regionorthe th o nt - west. Indeed, the evidence is in favour of a trans-peninsular route from Ballantrae Bay up the River Stinchar and its tributary the Duisk, thence three miles overland to the Cree, and it is likely that the Arran contacts seen at Cairnholy were established and maintained by this route rather tha coastiny nb g round Burro wMule th HeaGallowayf o ld dan .

LIS CHAMBEREP TO D TOMBS SHOW MAPN O N , FIG. 1 . 1. Newton, Anwoth. Remains of a rectangular chamber about 7 by 3 feet, with side slabs flush wit e groundhth thred an , e pillar-shaped stones fee5 inche4 to t p u scornerse higth t ha . Probably remain Clydea f so - Carlingford chamber. (Kirk. Invent., . 29.No ) 2. Cairnholy , KirkmabreckI . Present paper. 3. Cairnholy , KirkmabreckII . Present paper. 4. Boreland, Minnigaff. This cairn is virtually intact and without vegetation growing on it, and is 6 feet in height, ridged, and with a modern pyramidal cairn . it builFou n o tr e stonecrescentith f o s c forecourt settin visiblee gar e outermosth , t bot feeh4 inche6 t e s th high d an , innermost 1 foot 9 inches high. There is no sign of the passage or chamber, both being concealed by the cairn. There is no forecourt blocking visible e peristalitTh . s intermittentlhi y marke y largb d e boulders of the same-type as those of the facade. (Kirk. Invent., . 362.No ) 5. Dranandow, Minnigaff. Excavate . J H . . A Edward y b d n 1922i st bu , no finds. The cairn had been much robbed, and there was no built ker r peristalitho b . Fou e terminar on latera d an ll chamber, none having functional entrance. Vestigial septal slabs. (Kirk. Invent., No. 358; Edwards in Proc. Soc. Ant. Scot., vol. Ivii. (1923), p. 55.) . GrimeCf 1 Proc.n i s Prehist. Soc. (1939) lone . 139p , Th g . cair f Ringhano m Low, Derbyshire, shoy same wma th e process (Bateman Years'n Te , Diggings (1861) . 93)p , . 144 PROCEEDINGS OF THE SOCIETY, 1948-49.

6. Cairnderry, Minnigaff. Large roughly circular cairn with three radial chambers and a few stones of peristalith on the south-west. (Kirk. Invent., . 346.No ) 7. High Gillespie, Luced Ol . Large chambered cairn, pear-shapedt cu d an , across the narrow end by a field dyke. Traces of at least five, and probably originally six or seven, lateral chambers, and another on the axial line nea broae th r d end. (Wig. Invent., . 346.No ) 8. Mid-Gleniron Lucew Ne , , SmalII l chambered cairn, much robbedd an , destroye s reait rt a dend , wher a fiele d dyk crescentieA cut . it s c facade can be traced at the southern end, with indications of a heavy forecourt blocking t leas lateraa e d ton an , l chamber axiae Th l .chambe r visiblew no t .i sno (Wig. Invent., . 263.No ) 9. Mid-Gleniron I, New Luce. Large chambered cairn near no. 8, with half-facade surviving with high forecourt blocking, one axial chamber and one lateral, and a possible closed cist or chamber behind this. Cut throug y cart-trackb h ; some stone f peristalitho s . (Wig. Invent., No. 261.) 10. Caves of Kilhern, New Luce. Roughly pear-shaped cairn with axial chamber at wider end, two lateral chambers apparently without functional entrances secona d an ,d axial chamber nea e smalth r l end. hign I h upland country. (Wig. Invent., . 269.No ) 11. White Cairn, Bargrennan, Minnigaff. Present paper. 12. Auld Wife's Grave, Cairnscarrow, Inch. Megalithic burial chamber con- structed against a natural rock outcrop forming one side. It may be nothing more than a massive Bronze Age cist. (Wig. Invent., No. 46.) (Not visited.) 13. Cairn Kenny, High Murdonochee, New Luce. Circular chambered cairn of uncertain type, probably a passage-grave. (Wig. Invent., No. 271.) (Not visited.) . Cave14 Cairn, Strawarren, Ballantrae. Circular chambered cair uncertaif no n type, perhap passage-gravesa . Smith(J . , Prehistoric Ayrshiren i n Ma (1895) . 225.p , ) (Not visited.)

PART IV—EXCAVATIO THF NO E WHITE CAIRN, BARGRENNAN. The Site and its Setting (fig. 18). The Passage-Grave 1 knowe Whitth s a en Cairn, Bargrennane th s i , easternmost of two, sites of the same name, both north of Bargrennan, and a little over half a mile apart. The western White Cairn is more or less intact fee 0 n diamete5 i , t fee7 d t highan r , wit evidenco hn o t s a e t coverwhethei t tomb-chamberno a cissa r r o ro t . The eastern White Cairn lies withi e 375-foonth t contou a broa n o rd spur of land running south-west from Glencaird Hill (1025 feet). To the north-east is the great mass of The Merrick (2764 feet), separated from which by the Glen Trool valley lies Lamachan Hill to the east (2349 feet),

1 Nat. Grid Ref. 353784; O.S. 6-inch sheet Kirkcudbrightshire XXIII N.B., Royal Oomm. Auc. Mons. (Scot.), Kirkcudbrightshire Inventory (1914) . 350westere No ,Th . n White . 349CairNo . s ni EXCAVATIO THREF NO E NEOLITHIC CHAMBERED TOMBS5 14 .

WHITE CA/R-N

RUVMAWANTIE MODE.

JOO /3EO BARGRENNAN O 'A-

MIDDLE BRIDG CREF O E E

Pig. 18. Site plan of eastern White Cairn, Bargrennan.

VOL. LXXXIH. 10 146 PROCEEDINGS OF THE SOCIETY, 1948-49.

but the moorland immediately around the White Cairn is undulating, and not in general rising above about 450 feet. Southwards the view extends towards Luce mileBay8 north-wes1 e , th s awayo t d coase an , tth t Girvanta , 16 miles distant, lies behind a low range of hills rising to 700-800 feet. A mile cairsoute th f nho run Rivee sth r Cree vallea n i , y withi 150-fooe nth t contour. As at Cairnholy, the subsoil is Silurian rocks, but of the Wenlock series. Glacial deposits overlie the rock in most places, but there are frequent outcrops. The area is in general moorland and rough grazing, e easterth d nan White Cairlann o ds ni recently planteForestre th y db y Commission.

Visible Structures before Excavation (fig. 19). The site consiste a much-robbe f o d d circular cairn, wit ha diamete r of some 45 feet, remaining to a maximum height of 4 feet 6 inches near the centre but elsewhere averaging about 2 feet. Within this cairn was exposed a megalithic chamber and passage, of typologically late plan in which these two elements were not structurally distinct, the whole forming n "entranca e grave" narrowing toward s outeit s r feeend4 2 t, long, with its axis north-west and south-east. Two massive capstones still remained covering the inner half of the structure. There were no revetment stones or peristalith cair e partls th nwa d yan , covered with vegetation.

Method of Excavation. ** e vern vieth I yf w o ruine d nature monumentth f o e , attentios nwa directe excavatioe th o d t passage th f chambernd o examinatioe an th d an , n of an area at its presumed entrance, though no structural evidence of a junction betwee e passag nth e cair edgd th f enan eo coul seee db n before excavation plannes site ewa Th . d fro axian ma lfee0 scala lin1 tn f eo eo to 1 inch (1 : 120), with a detailed plan of the excavated area at 2 feet to 1 inch (1 : 24) and a longitudinal section at twice this scale (1 : 12).

Summary of Results. A collective chambered tomb, within the Passage-Grave group and an isolated member of its class in South-west Scotland, was built on a slope overlooking the Cree valley some 15-20 miles from the coast. Its builders not only followed architectural and ceremonial traditions alien to the buildere Clyde-Carlingforth f o s d tomb n thii s s region t thebu , y used potter a typ f o ye dissimilar from that foun t a Cairnholyd . Neae th r entranc tombe th o et , ritual makine , th includin d gan t diggine pi gth a f go in it of a fierce fire, followed by the deposition of a cremation and EXCAVATIO F THRENO E NEOLITHIC CHAMBERED TOMBS7 14 . a flint tool, suggests comparison with Late Neolithic practices known from elsewhere in Lowland Scotland and in Southern England. BAHGRENNAN WHITE CAIR.N

'

BUR.IAL CHAMBER.'

I t I I \

1O 15 2O 25

Pig. 19. DESCRIPTION. The Cairn. In the small area examined the cairn was found to consist of large rounded boulders, without evidenc fory builf an m o f eo t ker peristalitr bo h or larger blocks. 148 PROCEEDING E SOCIETYTH F O S , 1948-49.

The Passage d Chamberan (fig. 20). Since it is impossible to make any structural distinction between these two elements, it is convenient to describe the whole area between the orthostats and originally roofed as a single entity. The walls were con- structed of large boulders, set with their flat faces innermost, above which the roofing slabs had been carried on massive dry-stone walling. The structure took the form of a truncated wedge, 24 feet long, 4 feet wide internall innes it t yra (north-west) end narrowind ,an gopes tit feeo2 t na t (south-east) end. The walls were made up of eight stones, roughly paired, founs excavation anwa o dt d i n tha passage th t ceased eha exiso a dt s a t structural featur poina t ea t som feee8 t befor t reachei e e edge th dth f eo e south-westh cairn n O . t sid n additionaa e slaw lo bl appeare o cont d - tinu e linr anotheeth efo e are th a feet2 t r immediatelbu , fronn s yi wa t occupie e rituath t describepi y lb d d below barrieA . f largo r e stones, rather more compactly laid tha remaindee nth e cairnth f o ,r appeareo dt form a "blocking" across the line of the passage approach at the edge e cairnoth f , outside which wer smalo etw l burn ground ol te areadth n so surface t therBu e. could never have bee na functiona l entrance th o t e passage throug periphere hth cairne th accesd f yonlan o ,n sca y have been gaine descendiny db somn gi e manner e int caire edgth o th f n eo maso t s entrance th e beneat firse hth t capstone. Steps giving acces thin si s manner indeee ar d known fro mCotswola havy d ma tome d existeban d heret bu , wreckeo to aree s th retaio adwa t n evidence. The inner ten feet of the chamber were still roofed by massive cap- stone situ.n si These were carriereae th r y walling-stonedb very b yd an , heav walliny ydr f huggo e blocks parallelograa e feet2 on t , y se ,b 3 f mo on the orthostats at the sides. The under-side of the capstones were 4 feet 6 inches above the original floor-level, and it is likely that the roof was maintaine t thida s height throughou passagee th t . Corbel stoneo t s support the capstones still remained at one or two points along the passage, notabl north-easte th n yo . Osecone nth d south-wesstone th f eo t wall chambefroe reae th mf th ro r were various scratches, some forming rectilinear design anciente sb thay tma . The arewhole th af eo unde capstonee rth beed sha n thoroughly robbed, and only an inch or so of modern humic soil remained over the original floor in this part of the tomb. But thanks largely to the fact that the passage and outer part of what was presumably regarded the chamber were choked with debris, which included large corbel-slabs fallen in, it d beeha n left intact e falleTh n. slabs caused considerable difficultn i y excavating this confined area t eventuallbu , completels wa t yi y examined. It was found that the floor had been covered with a fairly regular paving of flat slabs, which had presumably originally extended over the CAP-5TONE. BARGREN'NAN -—~\J THE WHITE CAIRN

DETAIF O L CHAMBER. & PASSAGE

CAIRN 5TONE CORBEL.& 5 5

Ol Z 34- 5673 i i i i I I a==i FEET O 1 ^ ^ METRES CREMATED &OHES &, POT-JHERDJ OM PAVING

AREA BUR.NT RED '

'"'"JAMi!//>W r= ?£•. R.1TUAL CREMATION PIT P)™[%1{^ ik^ >-^...... -- ( \ fi

EDGE OF CAIRN

L/M/ EXCAVATIONF O T

Pig. 20. 150 PROCEEDINGS OF THE SOCIETY, 1948-49.

inner part of the chamber floor as well. On this paving, over an area extending for some five feet into the passage from the front edge of the second capstone, were scattered fragment f cremateo s d bone d sherdan s s of pottery ornamented with incisions, cord ornament and impressions probably made wit edge cockla h th f eo e (cardium) shell (fig. 21). Cremated bone fragments were also found scattered further towards the entrance, but without pottery e passagee fillinth Th f o .g , betwee e largth n e fallen slabs, consisted of stones and 'dark soil to a depth of 1 foot 6 inches to 2 feet.

The Ritual Cremation Pit. A completely unexpected discover s madywa e while searchine th r gfo continuation of the passage towards the edge of the cairn. Just off the

Pig. 21. White Oairn, Bargreunan: 1-4. Pottery from Passage; 5. Flint "fabricator' from Ritual Pit. (A)

axia structure l linth north-ease f eo th aren o ea t claf as o y wa tfloo r burnt intensn a o t e orange-red, abou fee4 t inche 6 feett3 toward d y an ,sb e sth north-eastern side of this was a clay-lined pit, 2 feet in diameter and a little ove foo1 r t deep, wit sides hit s similarly burnt. t leveFillinpi e l gth with the surrounding floor was a mass of black soil in which cremated bones were mixed with abundant fragments of oak charcoal, and on the thif o sp fillingto , nea north-wese th r calcinea t pite s edgth ,wa f eo d flint tool of the "fabricator" type (fig. 21, No. 5). e Parallelth n shali e t eas e w suco no finds t o se l yt t e hrituaa bu ar , t lpi subsequent discussion that something simila s beeha rn t founleasa n i dt one Passage-Grave (Bryn Celli Ddu) thad ,an t likely parallelse existh n i t Late Neolithic ritual hole t Cairnpappla s e Hil n Wesi l t Lothiand an , those from southern English sites such as Dorchester-on-Thames and Stoneheng. eI EXCAVATIO THREF NO E NEOLITHIC CHAMBERED TOMBS1 15 .

LIST OP FINDS. 1. Pottery. (EO. 843-45.) Sherds from paving passage.n i About 50-60 sherd f pottero s y representing at least three vessels. All fragments are of coarse ware, usually wite surfaceth h s flaked off d rangin,an n coloui g r from blackish brow red-buffo e pastnt th grie en i tTh include. s lumpf o s 0-2o t 5p u inch across therd an ,evidencs ei severan ei l sherd f conso - struction in overlapping rings or coils. No fragments of flat bases are present, but no other features of shape can be recognised. There are, however, 11 decorated sherds which can be divided into three groups: (a) Four sherd coarsf so e ware, non whicf eo whichf o join, o d htw an sho wa massiv e cordon with finger-tip impressions spacet da interval f abouo s interioe inch1 t Th . r surfaces have flaked away. Alon cordoe edge th gf on oblique o nar e incised lines, and two other sherds appear to continue these incisions in rough lattice pattern. A fifth sherd may be a fragment of the cordoned zone (fig. 21, Nos. 1 and 2). sherdso (bretainine Tw ) on , g both surface 0-d 5an s inch thick, with smala l raised cordon alon edge whicf gon eo h run horizontasa l line of fine cord impression. The other sherd has similar cord- impressed ornament in horizontal, vertical and oblique lines . 3) . No (fig, 21 . (c) Four sherd f coarso s e ware f whico , appeare hon como st e from nea roundee th r t 0-d9po basinca f eho thicksmoothee Th . d reddish outer surface has ornament of horizontal lines made either with some form of comb or more probably with the edge ocardiuma f . 4) shel . No l , (fig21 . This assemblage of pottery is difficult to match in Scotland, and problems it discussee sar d later.

2. Stone. (EO. 846.) "Fabricator" f whito e flint, burnt, 2-4 inche y 1-b s0 incy b h 0-5 inch maximum thickness. f burnFroo p mtto deposi rituan i t t pi l (fig. 21, No. 5). of this type are common in Bronze Age contexts, but are also known in Late Neolithic associations at, for instance, Dorchester-on- Thames and (Aubrey Holes). Its significance is discussed elsewher texte th .n ei

PART V—DISCUSSION OF THE BARGRENNAN EXCAVATION. Architecturally, the White Cairn presents so distinct a contrast to the Clyde-Carlingford tombs discussed earlie thin i r s paper tha t i must e b t considere monumene th s d a alie n a f no t grou peoplef po , sharing little save the basic fact of collective burial in a stone-built vault with their neighbours 152 PROCEEDINGS OF THE SOCIETY, 1948-49.

Cre e iLucd nth ean e valleyscirculas it n I .r cairn-pla undifferentiated nan d wedge-shaped burial chambe passagd an r onevirtualln s ei i t i , y impossible paralleo t knowe th n i ln Scottish tombPassage-Grave th f so e class. Typo- logically it bears a surprising resemblance to that homogeneous group of round chambered cairns one of us has defined in Southern Ireland and in western e CornwalIsle th f Scilly—tho sd an l e Scilly-Tramore group.1 Further than noting this close s wortsimilariti t i h t cannoe bu yw , go t while noting that other Passage-Grave f aberrano s t type o occud s n i r South-west Scotland, notably the double-chambered round cairn on the Water of Deugh,2 paralleled only in the cairn in Derbyshire. The pottery fro tome mth b emphasise s distinctioit s n fro e Clydemth - Carlingford series t beari s resemblanco a , sn froy Scottise an mth o et r ho Irish tombs of this culture, nor does it belong to the Hebridean or the Orcadian group of Scottish Neolithic wares. Its nearest parallels seem to Irise th hbn e i Passage-Grav e potter typee th f syo known from Lochcrew d Carrowkeelan , bot fabrin hi d ornamentan c , though eve- n re her e th e semblance doe t amounsno identityo t t : another possible analogue b y ema pottere foun Ronaldswae th th n di f yo y culture discussed below. The remarkabl entrance specifio th etomn e t s rituath a bha co t et pi l parallel in chambered tombs, so far as we are aware. The pit behind the chamber in the Bryn Celli Ddu Passage-Grave in Anglesey provides a partial analogy, but no more. But a comparison may be drawn between t Bargrennaa thi t pi s thosd nan f Lat eo e Neolithic date associated with cremation-burial t Cairnpappla s e Hill, West Lothian,a serie n f i o s d an 3 sites in Southern England, notably Dorchester-on-Thames and the first phase of Stonehenge,4 where, however, the pits are arranged in regular arcs of penannular rings. These pits, though associated with cremation deposit r neao rt n themselvesi themno e ar , s theo burntd yr containo , n burnt material thir fo s d phenomenoan , nearese nth t parallel seeme b o t s the pits in the cremation cemetery at Ballateare, Isle of Man, which is similarly of Late Neolithic date.5 The Ronaldsway culture, to which the Ballateare cemetery belongs, has certain links with chambered tomb cultures in, for instance, the incised stone plaques of Iberian affinities, e flinth t concave saws commo e Iristh hn n i Clyde-Carlingford tombsd an , the polished edge knives known not only from the culture, but from chambered tombs in Arran, and Caithness.6 Comparisons might indee made db e between Bargrennae somth f eo n sherds with rough incised lattice ornament potterd Ronaldswae an , th f yo y culture.

Powell in Proc. Prehist. Soc. (1941), p. 142. Proc. Soc. Ant. Scot., vol. Ixiv. (1929-30) . 273p , . Proc. Soc. Ant. Scot., vol. Ixxxii. (1948-9), p. 76. Atklnson and C. M. Piggott, Excavations at Dorchester, Oxon (1951). Proc. Prehist. Soc., vol. xiii. (1947), p. 161. Ibid., . 149p .

EXCAVATIO F THRENO E NEOLITHIC CHAMBERED TOMBS3 15 .

The flint tool, found burnt on top of the filling of the ritual pit, is of a type allie o thost d e called "fabricators nineteenth-century b " y anti- quaries. Such n faci tool to e d Latsoccuth en i rNeolithi c sitet a s Dorchester-on-Thame d Stonehengan s e already mentioned d survivan , e well into the Middle Bronze Age: they do not appear to be characteristic of the Ronaldsway stone . But one occurred in the Unstan chambered cair Orknen ni y (wit hpolishea d edge knife).

There seems then just a possibility that there1 may be some link between makere Bargrennae th th f so n Passage-Grav Ronaldswae th d ean y culture Isle oManf th feo mor,o n thoug o ed thae hw n indicate hypothesithia s sa s needing additional evidenc r confirmatiofo e r refutationo n t contacBu . t betwee Latn i nn e Gallowa NeolithiMa e Isl f th eo d o yan cn timey b s i s means unlikelye carinateth d an , d bowl from Cairnholfacn i e b ty ma yI connected with Manx rather than Ulster contacts, and it is worth noting e stonthath f ee o individuat axeth f o s l Ronaldsway type known outside e islandth e nearesth , t geographicall e thosyar e from Wigtownshiree on , from Stoneykirk, west of Luce Bay, coming very close to the Ronaldsway forms.2

APPENDIX A. THE " JADBITE " AXE-FRAGMENT FROM CAIRNHOL. YI The axe-fragment of "jadeite" from Cairnholy I has been discussed in general terms on p. 137, but the problems raised by its material are by no means simple. Dr W. Campbell Smith, Keeper of the Department of Mineralogy, British Museum (Natural History) s verha ,y kindly undertake n examinationa e petroth f o n- logical thin section of the Cairnholy fragment and compared it with the only other available thin section f axeo s s mad e jadeitf rocko th e f o se class from Britain. These thin sections, fiv numbern i e , have been e coursmadth f n ei eo the work of the Sub-Committee on the Petrological Examination of Stone Axes of the South-Western Group of Museums, and were placed at Dr Campbell Smith's disposal by kind permission of the Sub-Committee and its Petrologist, Dr F. S. Wallis. The numbering used in the following report is that of the Sub-Committee's records. Dr Campbell Smith point t thasou t proves "iha t d very difficul identifo t t y the mineral components of these 'jadeite' axes from the thin sections alone," and emphasises that his identifications are provisional, and may need revision. With these reservations repors hi , t run follows:s sa — "Cairnholy I (No. 563). A diopside-rock, texture allotriomorphic-granular. Grain-size e pyroxenabouTh t. s 0-colourlesi e mm 3 n thii s n section, biaxial, optically positive, with maximum extinction, y:. c 39° . Accessory minerals: granules of sphene, and very small areas of a colourless mineral (optically positive, uniaxial).

1 Proc. Soc. Ant. Scot., vol. xix. (1884-5), p. 351, fig. 16. 2 Proc. Prehist. Soc., vol. xiii. (1947), p. 148 n. 154 PROCEEDINGS OF THE SOCIETY, 1948-49.

Beaulieu, Hants (No .mai e 123)Th n .constituen t appear pyroxenea e b o st , probably diopside, but its crystals show a remarkably zoned structure. The outer shells of these crystals show pale green in thin section and are e definitelb weakl y ma y e pleochroicgreeax e n e ni stonth th f o o es , colour. Associated wit pyroxene hth whas ei possibls i t y zoisite alsd oan , grain epidotf so sphend diametern ei an smald . ean mm l garnet2 . o t p su This stone migh derivee tb d fro same mth e kin geologicaf do l environment as that from Cairnholy. Raftra, St Levan, Cornwall (No. 170). The section consists almost entirely omineraa f -likf o l fibroud ean s habit. Long colourless prisms form mose th e sectionr tpar th fo par f t o t consist i tbu , fibroua f so s masf so pale brown colour. Here and there clear areas show cross-sections of the mineral, but these possess no definite outline. The mineral has straight extinction, elongatio s positiveni , wit e optihth c axial angle very small. I believe this minera s sillimanitei l hav e (fibrolite).W ee . . herth . n ei Department a nice axe of sillimanite said to be from N.W. and another from Spain (Alham Granada)e d a t musI . t als notee ob d that sillimanit knows ei massivn ni e form from Scotland, Scratchbury Camp, Wilts (No. 289). Very fine-grained felted mass of prisms of actinolite associated with some patches of extremely fine-grained fibres pale brown in thin section and having straight extinction and positive elongation. This may be wollastonite (or sillimanite). In so far as nephrit varieta s ei f actinoliteyo s possibli t i , e tha name th t e coule db applied to the material of the axe, but one needs to see the specimen. Breamore, Hants (No. 300). Prisms of thickness 0-025 to 0-05 mm. forming rathea r felted mass. Extinction angl35°o t p ,e u biaxial , positive. Cross- sections show cleavages intersecting roughl t righa y t angles. This i s probably a pyroxene, and its properties are not incompatible with those ofjadeite. Preston, Weymouth, Dorset (No. 334). Very fine-graine. 0-o dt 2mm . 0-0 3mm granular texture. The main constituent here is a green actinolite, pale green in thin section and somewhat pleochroic; pale green for vibrations alon lengte prismse gth th f ho , colourles vibrationr sfo t righsa t angleo st this. Extinction angle 19°. Other mineral e occasionaar s l grainf o s epidote and pale pink zoisite (?), with lines of opaque 'ore,' possibly magnetite chemicaA . l test woul necessare db identifo yt y this opaque mineral wits Scratchbure A hth . y axe t mighi , permissible b t describo et e this materia nephrites a l t agai neede bu , n on se o st e .the stone. source th r e fo material e os "fth A axe-headse th f so canno,I t mak suggesy ean - tion for those from Scratchbury, Breamore and Preston, but for the Cairnholy specime nI should-gues s Brittan source th s supplyf ya eo . Possibly that from Beaulieu is related to the same type as that from Cairnholy, and if so, Brittany may be its home also. My suggestion that these types could be found in Brittany is only a guess, but I think it is a good one. " Some of the pyroxene-granulites (gneiss a pyroxene) of Brittany are known as 'jade Breton.' e rockTh s have been describe . LacroixA y b d . Mos f theso t e contain feldspar and/or scapolite and these minerals are not present in either the Cairnhol Beauliee th r yo u specimens."

Dr Campbell Smith adds that it would be desirable to make direct comparisons EXCAVATION OF THREE NEOLITHIC CHAMBERED TOMBS. 155 between Breton and British axes, and hopes that this may be possible at some future time. gazetteee Th follow8 Miso t 13 r . sp s Chitty'basie n hereo Th p c. lissma s i t Miss Chitty's work, with certain additions, including the new petrological identifica- tions discusse Dy db r Campbell Smith above e uncertaintievien th I f .w o f o s identificatio d gazetteean s p stressedha ma re e h nmus th ,f cours o t e onle yb regarded as provisional pending further petrological work on a large scale.

GAZETTEE F "JADEITERO " CEREMONIAL AXE BRITAINN I S ACCOMPANO T , P YMA . 138)p , (fig14 ..

Map Site. No. References. Museum. Notes.

ENGLAND. Cambs : Burwell Fen 28 Evans, A8I, p. 107, Ashmolean (Evans fig. 52. Coll.). Cambridge (Wark- 27 Fox, Arch. Camb. Cambridgd an . A e •• worth Street). . Beg.,10 . p E. Grantchester 29 B.M. (SturgeColl.). Histon 26 Fox, ACR, p. 10, Cambridgd an . eA pi. v. E. Triplow 31 VCH Cambs, vol. i. Cambridgd an . A e From barrow? p. 267. E. Cheshire : Lyme Park 20 Ant. Journ., vol. In possession of xii. p. 167. Lord Newton. Cornwall : Falmouth 42 Hencken, Arch. Truro. Corn., fig. 18o. Baftra, St Levan 41 Wellcome Hist Com.- b S.WSu - . Mus. (Marsden mittee, Thin Coll.). sectio . 170No n. No site 43 Nat. Mus. Antiq., Edinburgh. Derbyshire: Brierlow, Buxton 21 Evans, . ASI,pp Ashmolean (Evans 107-108. Coll.). Hopton, Wirks- 22 Ibid. worth. Devon : Bovey Tracey . 44 Proc. Devon. Arch. Torquay. Ex. Soc., I, Pt, 2, , 1433 . 74.Pt ; Dorset : Parkstone 38 Cambridge A. and E. 156 PROCEEDING E SOCIETYTH F O S , 1948-49.

Map Site. No. References. Museum. Notes.

ENGLAND. Preston, Wey- 45 Dorchester. Axe-amulet. mouth. 8.W. Sub-Com- mittee, Thin section No. 334. Sturminster 37 Farnham, Pitt- Marshall. Rivers Mus. Durham : Caldron Snout . 17 Proc. Soc. Ant. Newcastle,d 3r series, vol. iv. p. 199. Hants : Barton, Milton . 46 Proc. Soc. Ant., vol. •• xxi. p. 265; Proc. Hants F.C., vol. iv. p. 185. Beaulieu . 47 Winchester (Dale S.W. Sub-Com- Coll.). mittee, Thin sectio . 123No .n Breamore 39 Devizes Mws. Cat., Devizes. S.W. Sub-Com- , piii . . vapt . mittee, Thin section No. 300. Hill Head 40 Winchester (Mogg- ridge Coll.). Hordle 47 Proc. Hants F.C., Hartley Mus., . 21 . p vol . v . Southampton. Isle of Man : Glencrutchery . 18 of ListManx Manx Museum. Antiqs., (1900), p. 16. Kent : Canterbury 35 B.M. B.M. Guide (1926). p , 100, . fig94 . London : Mortlake . 33 Arch. Journ., vol. London Museum. Ixxxvi. 84 . p . Vauxhall . 32 Ibid. Ibid. Norfolk : South Lopham . 25 Evans, ASI, 107. p . Ashmolean (Evans Coll.) ?. Thetford . 24 B.M. (Sturge Coll.). EXCAVATION OF THREE NEOLITHIC CHAMBERED TOMBS. 157

Map Site. No. Beferenoes. Museum. Notes.

ENGLAND Suffolk: Wherstead, 30 B.M. Found uprighn i t Ipswich. soil, point upper- most. Wilts : ' ' Barrow near 34 Arch., vol. xliii. Stonehenge." . 406p ; Evans, ASI, p. 107. Scratchbury 36 Hoare, Ancient Devizes. S.W. Sub-Com- Camp. Wilts, vol. i. p. 70 ; mittee, Thin Cat. Stourhead section No. 289. . Coll.,11 . No

SCOTLAND. Berwickshire : Greenlaw . 9 Proc. Soc. Ant.Nat. Mus. Antiq. Site identified later Scot., vol. xxvi. Edinburgh. than map. p. 175. Caithness : No site 1 Anderson, Stone, Ibid. p. 338. Dumfries : Mains 15 Arch., vol. vii. p. •• 414; Evans, ASI, p. 108. Fife: Dunfermline 8 Evans, ASI, 109. p . Nat. Mus. Antiq. Edinburgh. Inverness : Daviot 3 Evans, ASI, p. 107. Ashmolean (Evans •• Coll.). Fort William (B. 4 Fort William •• Spean at). Museum. Kirkcudbright : Cairnholy I. 14 Present paper. Nat. Mus. Antiq. S.W. Sub-Com- Edinburgh. mittee, Thin section No. 563. Castle Dougla. s 16 Proc. Soc. Ant. Ibid. Scot., vol. xvii. p. 384. Perth: Aberfeldy 5 Evans, ASI, . 109p . Ibid. Battray . 6 Proc. Soc. Ant. Ibid. Scot., vol. xvii. p. 383 ; Ander- son, Stone, fig. 331. 158 PROCEEDINGS OF THE SOCIETY, 1948-49.

Map Site. No. References. Museum. Notes.

SCOTLAND. Roxburgh : Ounzierton 10 Proc. Soc. Ant.Nat. Mus. Antiq. Tw- o un axes e on , Scot., vol. xvii. Edinburgh. published. p. 383; Ander- sen, Stone, fig. 332. Stirling : Stirling 7 Ibid. Sutherland : Dunrobin 2 •• Dunrobin Oastle No locality given. region ? Museum. Wigtown : Glenjorrie, Glen- 12 Proc. Soc. Ant.Nat. Mus. Antiq. luce. Scot., vol. ix. p. Edinburgh. 356 ; Anderson, Stone, fig. 333. Glenjorrie, Glen- 13 Proc. Soc. Ant. luce. Scot., vol. Ixiv. p. 298. IRELAND. Donegal : Raymoghy 11 Nat. Mus. Dublin. Mayo : Tristia 19 Ibid. Tipperary : Nenagh 23 Ibid.

Addenda (not on map).

Site. References. Museum. Notes.

Lincolnshire : Isle of Axholme . Arch Journ., Private coll. Probably from vol. . ciii5 . p . Wroot. Perth: Cat. Nat. Mus. Nat. Mus. Antiq. Lochearnhead Antiq. (1892), Edinburgh. 32. EXCAVATIO F THRENO E NEOLITHIC CHAMBERED TOMBS9 15 .

APPENDIX B.

THE PITCHSTONE FRAGMENTS FROM CAIRNHOLY I. By F. W. ANDERSON, Geological Survey. I have sectioned the two larger pieces of rock from Cairnholy (i.e. that from behind the south portal stone and that associated with the carinated pot in the entrance blocking). All three (the third a small flake coming from under the clean earth spread in the forecourt) are undoubtedly pitchstone. I have compared them with material from Arran (the only pitchstone we have in our collection), and though the Cairnholy specimens are less devitrified I have no doubt that they come from the same source. The Scotstarvit specimen, like the majority of the Arran pitchstones, is more strongly devitrified. (Proc. Soc. Ant. Scot., vol. Ixxxii (1947-48), p. 263.) The rock from Dun Fionn, Arran, is, however, very nearly as glassCairnhole th s ya y specimens. Tabsolutele ob y certai woule non d nee examino dt e pitchstones from many sourcesface vien th i t f wt thao bu , t Arranearese th s ni t sourc thir efo s rockd an , of the very close resemblance between them, it would probably be a waste of time looking further afield.

APPENDIX C.

REPORE MOLLUSCTH N O T A FROM FORECOURT BLOCKING, CAIRNHOLY I. Dr A. C. Stephen, of the Royal Scottish Museum, kindly examined a typical sampl e deposith f eo f marin o t e mollusca fro e forecourmth t blockin Cairnn gi - reportd holan , yI s that four species were represented followine th n ,i g numbers:— Periwinkle (Littorina littorea) . 46 large, 83 small or medium. Limpet (Patella. 11 vulgata) . Mussel (Mytilus. 5 . edulis) Whelk (Buccinum undatum) . portion of one medium-sized animal.

APPENDI. XD

CHARCOAL FROM CAIRNHOLY I AND BARGRENNAN. By M. Y. ORE, Royal Botanic Garden, Edinburgh.

Cairnholy I.—(a) From forecourt blockin n south—extremelo g y friable fragment (Quercusk oa f so sp.hazed )an l (Corylus Avellana Linn.). (6) From stone-hol n forecourt—i e smalw fe a l fragmentf o s hazel (Corylus Avellana Linn.). Bargrennan.—From ritual cremation pit—fragments of oak (Quercus sp.). 160 PROCEEDING E SOCIETYTH F O S , 1948-49.

APPENDI. XE

E DISTRIBUTIOTH F BEACH-PEBBLNO E FLINT. The plano-convex knife from Cairnholy II, and probably that from Cairnholy I as "well, "were made of beach-pebble flint, as were those from the Arran chambered cairns—in both areas the only available source. Flint pebbles of up to 6 by 4 foune incheb n d sca sporadicall beace th n hyo below Cairnholy neee on d d an , e materialpresencw th ra loo t r e origifurtheo f flinkBu n ou th o e . f r tno fo r nodules on the west Scottish coasts raises the possibility that their distribution may in part be a function of the drifting of seaweed with such pebbles held in the roots, especially from such areas of naturally abundant flint as the Antrim coast. The Antrim flint fro Earle mth y Larnian sit t Campbeltownea , Kintyref o , Age, seems certainly a humanly carried import,1 and the same probably e applieflint-workersth o t s ' hoards from Portpatric n Wigtownshire,i k f o 2 uncertai t agaibu f Antrie no nag e carryin mth t flintBu g . capacit f certaiyo n of the seaweeds, notably Fucus vesiculosus and Ascophyllum nodosum (the Bladder and Knotted Wracks), which may drift long distances with relatively large stones hel theiy db r roots ignore,e discussiob y mus t an distributioe no tn d i th f no f no primitive man's raw materials in beach-pebble form.3 We are much indebted to Professor C. M. Yonge of Glasgow University, and Dr J. B. Tait, Senior Hydro- grapher to the Scottish Home Department, for their interested co-operation in this inquiry; while there is no doubt of the floating power of the seaweeds, Dr Tait reports -with regard to the present-day surface movements that '' resultant currents through the North Channel between Galloway and Antrim seem almost invariably to be directed northward, or a little to evidence th wese n th f northO o . I ethav . regarn . .ei currentso dt I , should scarcely expect flints to be transported from so far north in Ireland as Antrim into Wigtown Bay, or to Galloway generally; to , especiall e wesyth t sideindeed an , de wes righth tp u coast f Scotlando t , yes! And in lesser degree the Arran-Bute area might also be fed from Antrim, but the tendency is for currents rather to scour out the lower basin e Clydoth f e (below Arran) d carran , y everythin t betweegou n Kintyre and Northern Ireland, with a strong probability of deposition along the e wesmainlandth tn e islando coastth d n .o an s, Oceani cinfluence. s between Hea Malid dan n Head greatly enhance this tendency." s Cairnhola r Sofa s concernedyi , therefore y excludma e ew , seaweed drift a distributiofactoe sa th n i r t leasa f no t Antrim flint alon shoree gth , though this havy ma e affected Arra Butd nan e "an contributeo ds e sourceth o dt supplf so y available to the Clyde-Carlingford folk there. And in the wider context of primitive beach-combin r usefugfo l stones s worti t i , h -while quotin gpassaga e to which Symington Grieve drew attention. At the beginning of the nineteenth centur nativee yRadace th th f so k Archipelag mid-Pacifin oi c obtained iron from

1 Movius, Irish Stone (1942)e Ag . 177p , , with refs. 2 Proc. Soc. Ant. Scot., vol. li. (1916-7) . 121p , . Shore Yongea . 3Se M . e Cf. C (1949) Th , . 68p , ; Symington Griev Trans.n ei Hot. Soc. Edin., vol. xxx. (1929), pp. 72-103. EXCAVATION OF THREE NEOLITHIC CHAMBERED TOMBS. 161

cast-up wreckage d "receivean , , .i na simila r manner, another treasure, hard stones for whetting. They are sought for in the roots and hollows of the trees throwa se tha e stth up: irostoned nan s belonchiefse th whoo o gt t , m they must be delivered."1 The search for flint by the Neolithic colonists of Western Scotland cannot have been so dissimilar.

APPENDI. XF

ANALYSES OF SOIL SAMPLES FROM OLD GROUND SURFACE UNDER CAIRN, CAIRNHOL. YI

Analyses of soil samples representing the old land surface on which the cairn had been originally built were made at the West of Scotland Agricultural College Allarr byM d Johnso r Nicol D compared d an , nan d with samples taken froe mth modern surface south of the cairn. They agree that the soil under the cairn is most likely to have preserved the same composition as it had when originally built thereforn uponca d usee an ,e b d with some confidence soit particularla s lno i t I . y ric nutrientsn hi t mighbu , expectee b t beao t d r arabl yearw efe scropa r fo s without manuring under primitive farming conditions d woulan , d suppora t natural growt f scrub o hd woodland an e fielTh d . soil (outsid e cairnth e s i ) significantly higher in phosphate than that under the cairn, probably owing to modern manuring. The soil beneath the cairn is, on the other hand, higher in potash (query enrichment owing to plant ash after clearance by burning), but the results are not altogether conclusive on this point. This soil was also slightly more acid than the modern sample, but not acid enough to be infertile. The detailed resultfollows:s a e sar —

Loss on Available Available ignition Sample. Total P2O6. pH. P205. K20. (organic matter).

la. Soil below cairn, 187 mg./lOO g. 6 mg./lOO g. 6 mg./lOO g. 5-5 r 5-cent3pe . 2 inches deep, brown, with charred matter.

16. 7 inches deep, 188 4 4 5-5 4-2 „ yellow-brown, under la.

p soiIla n fieldi To l. , 300 5 4 6-1 13-7 „ 9 inches light grey loam.

116. Under 0 Il1 a 253 6 3 5-9 14*7 inches yellow- brown.

1 von Kotzebue, Voyage of Discovery into the South Seas (1821), vol. iii. p. 155.

VOL. LXXXIII. 11